Hugging Face's logo

datablations
/
lm1-misc-pile

TensorBoard
Model card Files Metrics Community
Muennighoff commited on
Commit
0f79083
•
1 Parent(s): 7d9b0e2

Add

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. 146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_PALM_prompt_0.json +1 -0
  2. 146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_PALM_prompt_1.json +1 -0
  3. 146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_PALM_prompt_2.json +1 -0
  4. 146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_PALM_prompt_3.json +1 -0
  5. 146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_PALM_prompt_4.json +1 -0
  6. 146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_PALM_prompt_5.json +1 -0
  7. 146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_explicit-graph-description2_0.json +1 -0
  8. 146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_explicit-graph-description2_1.json +1 -0
  9. 146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_explicit-graph-description2_2.json +1 -0
  10. 146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_explicit-graph-description2_3.json +1 -0
  11. 146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_explicit-graph-description2_4.json +1 -0
  12. 146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_explicit-graph-description2_5.json +1 -0
  13. 146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_implicit-graph-description_0.json +1 -0
  14. 146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_implicit-graph-description_1.json +1 -0
  15. 146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_implicit-graph-description_2.json +1 -0
  16. 146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_implicit-graph-description_3.json +1 -0
  17. 146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_implicit-graph-description_4.json +1 -0
  18. 146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_implicit-graph-description_5.json +1 -0
  19. 146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_non-explicit-description_0.json +1 -0
  20. 146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_non-explicit-description_1.json +1 -0
  21. 146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_non-explicit-description_2.json +1 -0
  22. 146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_non-explicit-description_3.json +1 -0
  23. 146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_non-explicit-description_4.json +1 -0
  24. 146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_non-explicit-description_5.json +1 -0
  25. 146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_very-explicit-description_0.json +1 -0
  26. 146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_very-explicit-description_1.json +1 -0
  27. 146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_very-explicit-description_2.json +1 -0
  28. 146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_very-explicit-description_3.json +1 -0
  29. 146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_very-explicit-description_4.json +1 -0
  30. 146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_very-explicit-description_5.json +1 -0
  31. 146m14b14b/eval/agg.lm1-146m-14b_GEM-wiki_lingua_en_article_summary_en_0.json +1 -0
  32. 146m14b14b/eval/agg.lm1-146m-14b_GEM-wiki_lingua_en_article_summary_en_1.json +1 -0
  33. 146m14b14b/eval/agg.lm1-146m-14b_GEM-wiki_lingua_en_article_summary_en_2.json +1 -0
  34. 146m14b14b/eval/agg.lm1-146m-14b_GEM-wiki_lingua_en_article_summary_en_3.json +1 -0
  35. 146m14b14b/eval/agg.lm1-146m-14b_GEM-wiki_lingua_en_article_summary_en_4.json +1 -0
  36. 146m14b14b/eval/agg.lm1-146m-14b_GEM-wiki_lingua_en_article_summary_en_5.json +1 -0
  37. 146m14b14b/eval/agg.lm1-146m-14b_GEM-wiki_lingua_en_rephrase_en_0.json +1 -0
  38. 146m14b14b/eval/agg.lm1-146m-14b_GEM-wiki_lingua_en_rephrase_en_1.json +1 -0
  39. 146m14b14b/eval/agg.lm1-146m-14b_GEM-wiki_lingua_en_rephrase_en_2.json +1 -0
  40. 146m14b14b/eval/agg.lm1-146m-14b_GEM-wiki_lingua_en_rephrase_en_3.json +1 -0
  41. 146m14b14b/eval/agg.lm1-146m-14b_GEM-wiki_lingua_en_rephrase_en_4.json +1 -0
  42. 146m14b14b/eval/agg.lm1-146m-14b_GEM-wiki_lingua_en_rephrase_en_5.json +1 -0
  43. 146m14b14b/eval/agg.lm1-146m-14b_GEM-wiki_lingua_en_summarize_above_en_0.json +1 -0
  44. 146m14b14b/eval/agg.lm1-146m-14b_GEM-wiki_lingua_en_summarize_above_en_1.json +1 -0
  45. 146m14b14b/eval/agg.lm1-146m-14b_GEM-wiki_lingua_en_summarize_above_en_2.json +1 -0
  46. 146m14b14b/eval/agg.lm1-146m-14b_GEM-wiki_lingua_en_summarize_above_en_3.json +1 -0
  47. 146m14b14b/eval/agg.lm1-146m-14b_GEM-wiki_lingua_en_summarize_above_en_4.json +1 -0
  48. 146m14b14b/eval/agg.lm1-146m-14b_GEM-wiki_lingua_en_summarize_above_en_5.json +1 -0
  49. 146m14b14b/eval/agg.lm1-146m-14b_GEM-wiki_lingua_en_tldr_en_0.json +1 -0
  50. 146m14b14b/eval/agg.lm1-146m-14b_GEM-wiki_lingua_en_tldr_en_1.json +1 -0
146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_PALM_prompt_0.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "bleu": 0.2304122240550486, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.02282542594810443}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_precision": 0.060241646379365375, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0018281076733804075}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_recall": 0.24592443008643936, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.00451257736641835}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_fmeasure": 0.088654136660333, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0018667826313676483}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_precision": 0.028611694237188352, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.001256141851839684}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_recall": 0.1211577510207888, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0030274510689492543}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_fmeasure": 0.04185850357277026, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0011315772112214053}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_precision": 0.059215633929813966, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0017815112040087317}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_recall": 0.2428343490294157, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.004471374721999319}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_fmeasure": 0.08728570338341435, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.001811977941125175}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_precision": 0.057472515102994094, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0017658517107930081}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_recall": 0.2343449114940915, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.004180785536759541}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_fmeasure": 0.08440847938326362, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0017495023319111394}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 0, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_PALM_prompt_1.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "bleu": 0.24093765005017717, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.0230030873247297}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_precision": 0.06608421860053589, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.002336019334031938}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_recall": 0.21220436916634114, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.004053789065265438}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_fmeasure": 0.08727381267513917, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.002019085465483018}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_precision": 0.026958060365485183, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0014325663154050872}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_recall": 0.09594460677384929, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.002636689924376356}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_fmeasure": 0.03690045692783099, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0011932344773442297}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_precision": 0.0627130320531275, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0022019609556712485}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_recall": 0.20593946270610541, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.003942128258116945}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_fmeasure": 0.08339382486023829, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0018750612270331736}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_precision": 0.06314633303729772, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.00225899090820857}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_recall": 0.20495711011670537, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.003873468183707786}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_fmeasure": 0.08340965215556247, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.001888288916759086}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 1, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_PALM_prompt_2.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "bleu": 0.233031239019028, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.018622989143777895}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_precision": 0.0669033461265895, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0022457356623633616}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_recall": 0.2210311245370995, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.004136964100813092}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_fmeasure": 0.0894179415607547, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0020679188927635165}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_precision": 0.02737441536683152, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0012163404235599126}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_recall": 0.10228061138551528, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0027171296640774193}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_fmeasure": 0.03847532772409176, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0011774128694059217}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_precision": 0.0627740873189878, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0020419307238977585}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_recall": 0.21409346975466498, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.004009271610351055}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_fmeasure": 0.08484682008716848, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0018654847610348828}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_precision": 0.06306585978755326, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.002078016863885293}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_recall": 0.21306042006535458, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.003959496879716326}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_fmeasure": 0.08486627231660553, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0018862109332949393}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 2, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_PALM_prompt_3.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "bleu": 0.23948204833652226, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.02987129954951102}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_precision": 0.06118912078285516, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0018874100598924987}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_recall": 0.21690381058073044, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.004037786222557515}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_fmeasure": 0.08516966035997384, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0019653410958603845}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_precision": 0.025691103456522035, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0009761929133606694}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_recall": 0.09913747741149387, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0026775514547381393}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_fmeasure": 0.036941223370484465, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0011451903532238477}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_precision": 0.057680973751972266, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.001710672833760511}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_recall": 0.21034092384649808, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.003897565257297013}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_fmeasure": 0.08113455126274569, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0017918361111978508}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_precision": 0.058077399704829655, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0017441871631557635}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_recall": 0.2102632726423538, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.003885092717560431}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_fmeasure": 0.08135704576559996, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.001809777621388229}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 3, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_PALM_prompt_4.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "bleu": 0.23249910495908727, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.029070914247746167}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_precision": 0.06371462011124131, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0019856815389907496}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_recall": 0.2217109306879308, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.003947138405833181}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_fmeasure": 0.08752577226961719, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.001988775968059262}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_precision": 0.026831790674440324, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.001147287521474585}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_recall": 0.10130367868890543, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0025744699017443726}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_fmeasure": 0.03752454514172853, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.001129399736645277}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_precision": 0.05966276178114979, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0017451674391315488}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_recall": 0.21514059835508997, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.0038265964816655085}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_fmeasure": 0.08315980831393469, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0017913382350644002}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_precision": 0.060216714540077834, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0017985198514999731}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_recall": 0.21496362989260864, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.0038038383785956613}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_fmeasure": 0.08355071339347263, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0018285792853000794}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 4, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_PALM_prompt_5.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "bleu": 0.22081286066426478, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.02274325142571731}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_precision": 0.06624132704477384, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.002139644596424952}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_recall": 0.22692456733410507, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.004091333470382997}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_fmeasure": 0.09147989131946721, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.002205390270555857}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_precision": 0.02768138829423821, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0011951077136256075}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_recall": 0.10373515662428917, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0027182295029531743}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_fmeasure": 0.03903350582230713, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.00122131545348737}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_precision": 0.06135390440151855, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.001830142897495819}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_recall": 0.2187629645915043, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.003916988813854803}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_fmeasure": 0.0861662328266173, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0019549257073660665}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_precision": 0.06154502173900534, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0018661465944842495}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_recall": 0.21792767487533674, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.0038849552847416537}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_fmeasure": 0.0861577572212862, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "3e41305c-5461-4cf3-853d-8a6fb5747623", "prompt_jinja": "I will verbalize an abstract representation of a sentence in natural language. To do so, I will first show the representation and then the natural language. The text needs to include all of the information in the representation.\n\n{{input | join(\", \")}} {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0019790861267728415}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 5, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_explicit-graph-description2_0.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_precision": 0.021212341576732085, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0007573290174140969}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_recall": 0.15394872315339908, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.0025907246130094292}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_fmeasure": 0.03406862468978302, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0007291397354522542}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_precision": 0.0007330327285282334, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.00013102839162176146}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_recall": 0.0060791270967935854, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0005096344336408659}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_fmeasure": 0.0010295704236990917, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 9.29161964969125e-05}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_precision": 0.020816161702046237, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0006635945167961655}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_recall": 0.15359448098622203, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.00258622832573184}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_fmeasure": 0.033775131528634185, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0006877591126027594}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_precision": 0.014879313282812858, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0006265155007215373}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_recall": 0.11208211458026511, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.001969123293960845}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_fmeasure": 0.023579045736592925, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0005251315581581948}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "bleu": 0.005621069280208802, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.00010070462717428731}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 0, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_explicit-graph-description2_1.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_precision": 0.1725932493840725, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.002529386481031639}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_recall": 0.1604306963623193, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.002484650213252913}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_fmeasure": 0.14612427698267197, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0018059192168654533}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_precision": 0.018139617142187602, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.001178810295076615}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_recall": 0.017838294270011764, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0010680407233547252}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_fmeasure": 0.015048493432535332, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0008510426237769514}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_precision": 0.1452889883810266, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0021500071787494027}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_recall": 0.13484344118305408, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.0020730900587942985}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_fmeasure": 0.12193728452169696, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0014342614701120422}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_precision": 0.1538719868843642, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0022759439069310527}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_recall": 0.1423996572873628, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.002170737056924197}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_fmeasure": 0.12940784544932987, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0015587985285077882}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "bleu": 0.7928900478653991, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.13053518516004906}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 1, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_explicit-graph-description2_2.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_precision": 0.21086754974453098, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.003481087853726294}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_recall": 0.25038766375814814, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.004322057797344624}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_fmeasure": 0.1907465620653806, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0025141994758677614}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_precision": 0.046020553876655636, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.002081339235092095}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_recall": 0.06454664623813919, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0024885004130050735}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_fmeasure": 0.04229150796312311, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0016292319117673752}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_precision": 0.17652392027199135, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.003026510309888636}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_recall": 0.2138167297173281, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.003888991530572478}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_fmeasure": 0.15945985445699104, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.00211755282820529}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_precision": 0.18652878286693028, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.003173835353071881}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_recall": 0.21970003626695772, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.003760387533497808}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_fmeasure": 0.16758058695841305, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.002234265089238999}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "bleu": 1.035688482397717, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.07979180056283594}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 2, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_explicit-graph-description2_3.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_precision": 0.24945289061580536, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.004234785605850604}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_recall": 0.27695089051506727, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.004409164518913342}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_fmeasure": 0.2155827425525286, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0028813429865885256}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_precision": 0.06789221743946505, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.002570766391550549}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_recall": 0.08374766031914743, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.002770294903947266}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_fmeasure": 0.05870693940495815, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0019495378220205084}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_precision": 0.20844479328303336, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0036153889991536138}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_recall": 0.2362160338231045, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.003930516775517815}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_fmeasure": 0.1799141924288802, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0023927111472121488}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_precision": 0.21975278899164405, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0038037794972502884}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_recall": 0.24068798176788908, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.0037486405001691237}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_fmeasure": 0.18810543014252795, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.00247964907125972}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "bleu": 1.4760612258997623, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.1344204674137375}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 3, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_explicit-graph-description2_4.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_precision": 0.2843113295309021, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.004668202372322337}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_recall": 0.2840533638331491, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.004212420357643946}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_fmeasure": 0.23277564344319238, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.003033324883659851}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_precision": 0.0855143860050401, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0028864190680247532}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_recall": 0.09058538636969998, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.002732539141359347}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_fmeasure": 0.06941913517884618, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0021154296895269154}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_precision": 0.24061329765622666, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0040548049932821436}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_recall": 0.2451732467444115, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.003800935008924143}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_fmeasure": 0.1969438842373623, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.002591786476274177}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_precision": 0.2513996234316726, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.004211644541437699}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_recall": 0.24901463870097312, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.0036333762169696126}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_fmeasure": 0.20445884584732335, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.002660931367820124}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "bleu": 1.4525772102533077, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.09991573236218197}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 4, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_explicit-graph-description2_5.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_precision": 0.3079953699411845, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0052262700820691855}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_recall": 0.2990488948523553, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.004288889844947112}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_fmeasure": 0.24428760561857069, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0032770994487209613}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_precision": 0.09878776942732348, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.003295722862390014}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_recall": 0.09931014156182742, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.00285802935368016}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_fmeasure": 0.07628545628688663, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0022936944399277277}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_precision": 0.26091647796513057, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.004518606930149643}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_recall": 0.25924988114075476, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.003916835233501827}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_fmeasure": 0.20682176993182852, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.002772758722894698}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_precision": 0.2723426426961616, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.004751841751952048}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_recall": 0.26102854194007674, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.0037065483496949393}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_fmeasure": 0.21367093331703443, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0028579521145505734}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "bleu": 1.5535486678139987, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "afeec167-f75f-4687-a775-1efde7d04780", "prompt_jinja": "{{input | join(\", \")}}. \n\nThe above is a set of subject | predicate | object expressions separated by commas: \nWrite all the information in proper sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.12636178897739148}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 5, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_implicit-graph-description_0.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "bleu": 0.06678968022431767, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.013291302987332878}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_precision": 0.029914442023692294, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0013899685699484238}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_recall": 0.17191853554387962, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.0027258688235086617}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_fmeasure": 0.04180906676802502, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.000892515049715985}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_precision": 0.002770987452397337, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.00026930245436038245}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_recall": 0.018458908430645037, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0012055847911475867}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_fmeasure": 0.0043358346468287505, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0003699728130996654}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_precision": 0.028773567135668157, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0012807801393214825}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_recall": 0.16962928822415665, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.0026788882785198187}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_fmeasure": 0.04069689961336112, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0008129334087216642}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_precision": 0.023429395403589064, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0012857906320802946}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_recall": 0.13119208466933271, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.0020710838890581453}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_fmeasure": 0.03168737424504007, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0007282033005749001}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 0, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_implicit-graph-description_1.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "bleu": 0.7343903998488392, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.1264580334548859}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_precision": 0.1713694988441509, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0025683178287328103}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_recall": 0.1620927633038664, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.0026004642336094302}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_fmeasure": 0.1451954506401401, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0018476593880312048}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_precision": 0.018149248560522465, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0011980307264669568}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_recall": 0.018701683687898776, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0011131166018444143}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_fmeasure": 0.015147654240964752, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0008531589600636162}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_precision": 0.1438174575375414, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.002179050759564507}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_recall": 0.13648488466088793, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.002207259216754237}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_fmeasure": 0.12085106733704139, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0014628495715754606}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_precision": 0.1522841828379344, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.002306578353042297}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_recall": 0.14316734345889626, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.002243661271505656}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_fmeasure": 0.12811642696509545, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0015874918061787038}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 1, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_implicit-graph-description_2.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "bleu": 0.8950469728314032, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.06813864403493262}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_precision": 0.2245739569864698, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.00394768209564399}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_recall": 0.2605817818040319, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.004397225408663707}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_fmeasure": 0.19598922586079662, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.002726474308866776}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_precision": 0.05842649937474419, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0025228225800757995}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_recall": 0.07857813168613167, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.002700572476720445}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_fmeasure": 0.05131564722525416, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0018242084011574487}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_precision": 0.19008397370997976, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.003446693936021226}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_recall": 0.2280794378801942, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.004143086530714078}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_fmeasure": 0.16612890671250619, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.002321135524749462}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_precision": 0.19875997968955347, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.003614311300045863}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_recall": 0.22698955916927938, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.0037471810061475252}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_fmeasure": 0.17165207996551565, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0024000508163840063}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 2, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_implicit-graph-description_3.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "bleu": 1.0328999153691525, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.15218614003775394}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_precision": 0.2561698960860146, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.004586388937239763}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_recall": 0.2889121821334627, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.004530658302121611}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_fmeasure": 0.21578406129162228, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.003014899947915645}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_precision": 0.0756534526059316, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0028609493658018873}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_recall": 0.09619008514491978, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0029113141102096165}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_fmeasure": 0.06387420889427227, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0020202475334694754}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_precision": 0.2162947950404591, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.003924794592166209}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_recall": 0.25398159222153943, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.004265191984478587}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_fmeasure": 0.1831681982588611, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0025326526566856545}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_precision": 0.22494600706563603, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.00410385917148987}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_recall": 0.25165737151714396, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.0038988712792617612}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_fmeasure": 0.18805417860233245, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0026019608331974494}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 3, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_implicit-graph-description_4.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "bleu": 1.1848438797934058, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.10113709897455576}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_precision": 0.29191494117074934, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0048753981466979806}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_recall": 0.2982482419805292, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.004362410145983633}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_fmeasure": 0.23507855170285297, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.003052960609273728}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_precision": 0.09264180506664611, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0030441647048234763}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_recall": 0.10366534795774687, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.002886282274864994}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_fmeasure": 0.07429307318653183, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.00210574121245154}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_precision": 0.24944159316145484, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.004252142505917989}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_recall": 0.26299891956132837, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.004120440439329194}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_fmeasure": 0.20109163102106464, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.002579325282186586}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_precision": 0.25743050877782325, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.004415034565833493}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_recall": 0.25969769735732884, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.003739952618361054}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_fmeasure": 0.2054154975263721, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0026590981957875476}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 4, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_implicit-graph-description_5.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "bleu": 1.5291942430627856, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.12407405811462165}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_precision": 0.31817552785971276, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.005410918975610136}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_recall": 0.3120825384968243, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.004423006280320729}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_fmeasure": 0.25074445308374155, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.003280258168301159}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_precision": 0.11020791501317337, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0037693744168018546}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_recall": 0.114958415897581, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.003095157786907941}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_fmeasure": 0.08476444821349621, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.002405571742153227}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_precision": 0.2709202785200124, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.004714816857530811}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_recall": 0.2753613124108194, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.004151926334002586}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_fmeasure": 0.21469809323475145, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.002784872235569171}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_precision": 0.28086204710118207, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.004933792503425808}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_recall": 0.27229001474697245, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.003806628830915923}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_fmeasure": 0.21938893610824683, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "38342608-5cd7-4ce7-b2e1-905ecd7f4c80", "prompt_jinja": "{{input | join(\"; \")}}\nThe above is a collection of relations. Write descriptive English that contains this information.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0028754942341362322}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 5, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_non-explicit-description_0.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_precision": 0.048154071638123803, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0017394492115037211}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_recall": 0.1603960438769496, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.002808863528216255}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_fmeasure": 0.054872501275754215, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0011675161117053734}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_precision": 0.0050437692571547055, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.00041097690347038956}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_recall": 0.020874718274915008, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.00145782883855474}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_fmeasure": 0.00644935707690507, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.00048277790299858106}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_precision": 0.044143178163838545, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0015820050437832095}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_recall": 0.14857144649500334, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.0026359059490226556}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_fmeasure": 0.05035459707061074, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0010577917832862004}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_precision": 0.04435825561436467, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0016624296413482802}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_recall": 0.14581597687921308, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.0024815309695857325}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_fmeasure": 0.04968291242100845, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0010556937198168178}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "bleu": 0.06942821394863977, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.00721092856708907}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 0, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_non-explicit-description_1.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_precision": 0.16455416189125624, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0025710885946127505}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_recall": 0.2181617130834446, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.004232534637085028}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_fmeasure": 0.15006174394417673, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0019223053607641332}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_precision": 0.023103353875520607, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0012429378612023923}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_recall": 0.051688701593469964, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0024769333300300395}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_fmeasure": 0.023945131921722818, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.001073921369797315}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_precision": 0.13925189436630736, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0021920195516649307}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_recall": 0.19238762602768703, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.0040704144827113}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_fmeasure": 0.12712565779418603, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0015763640466711865}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_precision": 0.14613811038014662, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0023432542987913535}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_recall": 0.19219266107645588, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.0036970487118417475}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_fmeasure": 0.13222747966833925, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.001685437840262139}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "bleu": 0.7207771490116592, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.10258109032487232}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 1, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_non-explicit-description_2.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_precision": 0.1840518032934248, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0033695422988167393}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_recall": 0.32294580089178854, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.005293640122083205}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_fmeasure": 0.18077149568804796, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0024357765771144814}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_precision": 0.048339216717895615, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0021060002829076923}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_recall": 0.11482202688362075, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0034238195259001535}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_fmeasure": 0.050683857012723176, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0016201788737892754}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_precision": 0.1584513791779987, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.002944143746092905}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_recall": 0.2930866134885274, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.005175150990392834}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_fmeasure": 0.15676881729864572, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.002052846994821611}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_precision": 0.163518547765267, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.003098199168660185}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_recall": 0.28299351899056346, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.004588174178628674}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_fmeasure": 0.15894468731506023, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0021537616662293205}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "bleu": 0.8172765885695147, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.050751327067539105}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 2, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_non-explicit-description_3.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_precision": 0.21339019477940072, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.004136854182002759}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_recall": 0.36172056828283894, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.005192684637361719}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_fmeasure": 0.20493909153322515, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.002712950527519467}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_precision": 0.06875941454459712, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0027123104753735013}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_recall": 0.14026102787918196, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0035390177764360073}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_fmeasure": 0.06763126769213473, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0019242155202766913}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_precision": 0.18497245590804035, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0036161696226493217}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_recall": 0.32950137116051886, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.00509054750559144}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_fmeasure": 0.178978835556707, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0023052454882151385}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_precision": 0.18870247477597205, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0037949931460892973}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_recall": 0.31437773260575974, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.004468573983043156}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_fmeasure": 0.17901076394504797, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0023948023589502633}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "bleu": 0.8232200037553864, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.0990062331795447}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 3, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_non-explicit-description_4.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_precision": 0.24073524945767655, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.004707514658959234}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_recall": 0.37222777341340585, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.005100896049800429}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_fmeasure": 0.2219276041114797, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.003043064081835907}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_precision": 0.08343573004395873, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0029901511610586507}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_recall": 0.1518671437094746, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0036105628962266956}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_fmeasure": 0.07863230130745369, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.002070687205307558}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_precision": 0.2082283461845528, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.004101899636825203}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_recall": 0.3382840179138503, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.005009409498199744}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_fmeasure": 0.19296499339169593, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0025313363149249185}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_precision": 0.21271916694519305, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.00429809264467102}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_recall": 0.32505704280562625, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.004417521948317753}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_fmeasure": 0.19372279394640554, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0026569278397281402}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "bleu": 0.9445938185251443, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.09197403951561334}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 4, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_non-explicit-description_5.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_precision": 0.26082598406654584, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0051359481387504655}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_recall": 0.3729649784159858, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.004970827450327891}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_fmeasure": 0.23144425717217612, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0031085174926612394}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_precision": 0.09576839794606369, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0035521748356882993}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_recall": 0.1532169603638125, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.003593558920823611}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_fmeasure": 0.08355597806644839, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.002189706487328105}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_precision": 0.22583373330234002, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.004534237669240004}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_recall": 0.3374628358280029, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.004847122676128511}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_fmeasure": 0.2012929842528386, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.002610242315697286}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_precision": 0.23006368905799984, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.00470015880813854}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_recall": 0.3255358190359428, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.004336770681537931}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_fmeasure": 0.2020118383660613, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.002720953194833289}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "bleu": 1.0828952105973466, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "9415bd8a-685f-4fa4-803a-f09bd47d4603", "prompt_jinja": "I am taking this tabular data, where each row is separated by a \"|\" and expresses a relation between an object and a predicate : {{input | join(\", \")}}. \n\nNow, I will produce a description of the tabular data using English sentences. {% for i in references %}\n ||| {{ i }} \n{% endfor %}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.08428338494461286}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 5, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_very-explicit-description_0.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_precision": 0.09839982038939207, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0016034845059760892}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_recall": 0.557046636180574, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.004459443954469606}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_fmeasure": 0.15932996809565791, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0021975666324784045}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_precision": 0.03974822393863394, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0009259214471201931}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_recall": 0.23724700480115124, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.003530724556191307}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_fmeasure": 0.0645288094675153, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0013271654890972852}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_precision": 0.08556676553539082, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0012344258273556512}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_recall": 0.5120237901085162, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.004389079693859156}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_fmeasure": 0.13994234816561368, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.001713279737012679}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_precision": 0.08372529560944662, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.001428553657612201}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_recall": 0.47520736805531627, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.003928372595420955}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_fmeasure": 0.13548370931568984, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.001965481361266326}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "bleu": 0.70556499870502, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.07021163423422254}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 0, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_very-explicit-description_1.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_precision": 0.15192244328141744, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0026851175239504146}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_recall": 0.4088076221119547, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.005307206147263785}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_fmeasure": 0.1795535626460401, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0023078723613743383}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_precision": 0.042264113746990004, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.001519132737951407}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_recall": 0.15816170460414825, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.003648850558319156}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_fmeasure": 0.056216955504673816, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0015622457935091263}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_precision": 0.13222718563169758, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0023137442474583567}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_recall": 0.37772895967872394, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.005250016280580306}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_fmeasure": 0.15827887417485154, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0019506890461920023}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_precision": 0.1331053325871554, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0024745486155965113}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_recall": 0.3552035560328879, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.004588945170502327}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_fmeasure": 0.15584864004949936, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0020620407818320073}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "bleu": 0.7666335106831184, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.0635865012850457}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 1, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_very-explicit-description_2.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_precision": 0.18082568461374282, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.003639459051365589}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_recall": 0.4264948458628402, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.004935119334379477}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_fmeasure": 0.1995262410573067, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0024883401069483438}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_precision": 0.06212406576427143, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0023869401244636375}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_recall": 0.17467748933722366, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.003536521311814631}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_fmeasure": 0.07054805798392726, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0016830190110589496}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_precision": 0.15912923447907845, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0032345958330729694}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_recall": 0.39493138391122135, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.004915557812492545}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_fmeasure": 0.1770227263859978, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0020868398943825295}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_precision": 0.15978841140185437, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0033957422368174534}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_recall": 0.37244471989166417, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.004290031616626638}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_fmeasure": 0.1741607140832361, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0022309499919646618}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "bleu": 0.8036503769411473, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.08352144767185858}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 2, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_very-explicit-description_3.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_precision": 0.20161323138180814, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.00419614102063665}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_recall": 0.4284437489377629, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.004747351313475394}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_fmeasure": 0.20998314436048413, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0026958272587435566}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_precision": 0.07075454828433567, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0025432659534373814}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_recall": 0.17982051325515352, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0035838481085783518}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_fmeasure": 0.07655162918992574, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.001828909076942692}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_precision": 0.17689528139425809, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0036676387595752816}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_recall": 0.39708608920535615, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.004751490642713908}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_fmeasure": 0.1863768106489585, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.002259606335187374}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_precision": 0.1782191584681922, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0038632898410200503}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_recall": 0.37361714826137415, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.004151105270083309}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_fmeasure": 0.18356302820432102, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0024163227573974415}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "bleu": 0.9213985772786043, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.09261357714373215}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 3, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_very-explicit-description_4.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_precision": 0.22590230879356132, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.004885739171030689}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_recall": 0.4183617818367607, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.004646737629595512}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_fmeasure": 0.2203882766391892, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.003092911335667608}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_precision": 0.08369577119049203, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0029217645456625545}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_recall": 0.17612573754542205, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0035387797965567075}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_fmeasure": 0.08291524084048744, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.001997647183245828}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_precision": 0.19700967528722899, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.00420918433484521}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_recall": 0.3860053232564878, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.004602487402565482}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_fmeasure": 0.19451918196713988, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0025549062189435443}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_precision": 0.19774103481256436, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.004390815504231839}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_recall": 0.3641519676371665, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.004085150565177179}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_fmeasure": 0.1912941284472393, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.00269200231433118}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "bleu": 0.8669579841221637, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.08180503550292748}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 4, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-web_nlg_en_very-explicit-description_5.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_precision": 0.2333099109106658, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.004918167379356697}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_recall": 0.40468387245463144, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.004615290964853879}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_fmeasure": 0.221868544853033, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.003062043218874345}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_precision": 0.08575749645339896, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0029475165939869268}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_recall": 0.16978922632954452, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.00346238061542852}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_fmeasure": 0.08265916020277501, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.001971954733869372}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_precision": 0.20506533085409256, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.004283642129477899}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_recall": 0.37349518444433055, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.004568965644534639}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_fmeasure": 0.19660756094703621, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0025597130950611555}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_precision": 0.20613211140017276, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.004444626909561573}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_recall": 0.3539709992804558, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.004068457607282714}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_fmeasure": 0.19418171975741957, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.002689265498940047}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "bleu": 0.949825770028968, "fixed_answer_choice_list": null, "dataset_path": "GEM/web_nlg", "dataset_name": "en", "subset": null, "prompt_id": "426b682e-e801-4e8d-9ac3-5b676c9d3da2", "prompt_jinja": "A semantic triple is the atomic data entity in the Resource Description Framework (RDF) data model. As its name indicates, a triple is a set of three entities that codifies a statement about semantic data in the form of subject\u2013predicate\u2013object expressions. (e.g., \"Bob | is | 35\", or \"Bob | knows | John\"). \n\nA graph can be formed from a set of these triples. An example is {{input | join(\", \")}}. \n\nWrite grammatical text expressing all the relations succinctly and fluently.\n{% for i in references %}\n ||| {{ i }} \n{% endfor %}\n\n", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.07260355364195568}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 5, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-wiki_lingua_en_article_summary_en_0.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_precision": 0.1975977152505175, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0019210439152713556}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_recall": 0.3571450155403217, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.0027159219787335877}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_fmeasure": 0.23668156796351772, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0017834284452177681}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_precision": 0.04679995732393755, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0008453000647757492}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_recall": 0.08872427020642143, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.001757359421073499}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_fmeasure": 0.056425445419599346, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0009610968832196708}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_precision": 0.13079025220025553, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.001163188240810311}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_recall": 0.24818548845368613, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.002206255979336452}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_fmeasure": 0.15899265084812766, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.001137037008281727}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_precision": 0.18296614395904412, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.001771259927022808}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_recall": 0.3318205839417274, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.0025527759940675494}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_fmeasure": 0.21937178453510964, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0016467565432194853}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "bleu": 2.526005202191506, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.058273418733684156}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 0, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-wiki_lingua_en_article_summary_en_1.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_precision": 0.15054373183476955, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.001998147829744307}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_recall": 0.22938700618092542, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.003041552612274185}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_fmeasure": 0.1625536760242897, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0019424044371032676}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_precision": 0.026747967720227677, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0008081761640449396}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_recall": 0.04688223062252883, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.001502981154526947}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_fmeasure": 0.03048470823212383, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0008439966461249648}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_precision": 0.11130833909157777, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.001381727326473868}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_recall": 0.17251099505438725, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.002267024463807693}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_fmeasure": 0.11990114200505439, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0012809983090350478}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_precision": 0.14096779097971843, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0018582505585400385}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_recall": 0.21419795786755574, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.00281418674755583}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_fmeasure": 0.15185020212271533, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0017871174330384631}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "bleu": 1.7313590804890209, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.05908907418295776}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 1, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-wiki_lingua_en_article_summary_en_2.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_precision": 0.14664845124249912, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0021074927136243747}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_recall": 0.22029309843953365, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.002950858143159719}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_fmeasure": 0.15659972337101613, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.001906663384031481}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_precision": 0.025832066720273383, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0008715375341951669}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_recall": 0.04297038579858274, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0014465923870447984}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_fmeasure": 0.028283684834043706, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.000821620933532687}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_precision": 0.11186134311345497, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0015665677693472357}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_recall": 0.17054780622395707, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.002242184695689906}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_fmeasure": 0.11924857854706734, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0013235141134357415}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_precision": 0.1363376415997164, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0019672230168232964}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_recall": 0.20473556515895666, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.002733779107620804}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_fmeasure": 0.14532637028195308, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0017544389892467619}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "bleu": 1.5124106778411996, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.06141046632956911}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 2, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-wiki_lingua_en_article_summary_en_3.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_precision": 0.1237560514705609, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.002282039852755364}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_recall": 0.17609270530427382, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.0030210209846962413}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_fmeasure": 0.1258659396327009, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.001992963699044068}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_precision": 0.020954462249452663, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0007965730554829935}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_recall": 0.033657403446598254, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.001329371810945043}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_fmeasure": 0.021957452841427055, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0007360198417813418}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_precision": 0.09658445599443373, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.001753215991240076}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_recall": 0.13975525526667698, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.0023736809100178683}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_fmeasure": 0.0981193897316493, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0014598347738579937}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_precision": 0.11453614366434078, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.002105151678823837}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_recall": 0.16338969575022716, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.0028019873925272075}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_fmeasure": 0.11650869217135837, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.001832893857886487}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "bleu": 1.3802963074575254, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.06732152627137003}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 3, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-wiki_lingua_en_article_summary_en_4.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_precision": 0.04395368484664138, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0018424926390462097}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_recall": 0.05774517711149188, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.002305158478827759}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_fmeasure": 0.04204648300277925, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0016015557898894717}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_precision": 0.00844841454877157, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0006603987371609425}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_recall": 0.012177055063554206, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0009005039320534978}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_fmeasure": 0.008202499844200142, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0005608948554894471}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_precision": 0.035216144797083765, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0014914106117576477}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_recall": 0.04621071195954536, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.001796318755601515}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_fmeasure": 0.03339087455481403, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0012365521644087543}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_precision": 0.04090439759741698, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0017246420675129577}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_recall": 0.05360227908215058, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.0021376470771195598}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_fmeasure": 0.0389854987222177, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.001481057274442113}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "bleu": 0.22377593301174895, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.022223303535369093}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 4, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-wiki_lingua_en_article_summary_en_5.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_precision": 0.006968711039755044, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0008981875218117217}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_recall": 0.009045805133122158, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.0010585291219916539}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_fmeasure": 0.00624650026917102, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0006896033744358352}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_precision": 0.00147397688199298, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.00031382142309567925}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_recall": 0.0023038793256885807, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0005038053110111991}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_fmeasure": 0.001354763281978201, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0002354570603752033}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_precision": 0.005684552101183715, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0007811869359473511}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_recall": 0.0073860095468518735, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.0008694725694459097}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_fmeasure": 0.004998030899560508, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0005484876409153318}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_precision": 0.006504958900325744, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0008576188233041718}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_recall": 0.008412531590010163, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.0009903950820560048}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_fmeasure": 0.005785170257262004, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.000640536066601862}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "bleu": 1.3197166417911808e-08, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "2038df7b-5420-4a33-87ec-09715419deef", "prompt_jinja": "Article in English: {{source}}\n\nSummary in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "bleu_stderr": 2.839866989657345e-08}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 5, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-wiki_lingua_en_rephrase_en_0.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_precision": 0.09557902324303133, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0015934233011554208}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_recall": 0.15195533247270118, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.002167076868910249}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_fmeasure": 0.10814475341505056, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.001524767202301828}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_precision": 0.010021499848992905, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.00042254787804722635}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_recall": 0.0175643734135993, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0008555885514606594}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_fmeasure": 0.011588030659270141, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0004784186900067079}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_precision": 0.08020042698947122, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0012460897241632469}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_recall": 0.13117713414918547, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.0018429649355817298}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_fmeasure": 0.09167513906886186, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0012095215627801854}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_precision": 0.08997576030672756, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0014864753127451467}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_recall": 0.1439070530072875, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.002057912841397234}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_fmeasure": 0.10198753546847392, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0014211405382431829}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "bleu": 0.6145809763628965, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.05797766316597433}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 0, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-wiki_lingua_en_rephrase_en_1.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_precision": 0.11274337869180198, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0015080376244132586}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_recall": 0.10983840856878206, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.0014551714150494175}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_fmeasure": 0.0979457603122117, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0011364196004116048}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_precision": 0.004607065202297954, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.00029696477370218157}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_recall": 0.004443770455530801, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.00030921576747671716}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_fmeasure": 0.003914558098208722, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.00023624265329673873}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_precision": 0.09101884007968876, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.001195226939453244}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_recall": 0.08894440577554404, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.001162325195438461}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_fmeasure": 0.07861266088907197, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0008536301328765664}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_precision": 0.10885553700903176, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0014410450568103732}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_recall": 0.10635049542751197, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.0014076769390450833}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_fmeasure": 0.09467060993519806, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0010886015566558735}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "bleu": 0.2408792681226857, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.024130354527193396}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 1, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-wiki_lingua_en_rephrase_en_2.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_precision": 0.11450203486111951, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0015512493201278732}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_recall": 0.11764026218464517, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.0017763168006867999}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_fmeasure": 0.10129546926820872, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0012376127878624412}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_precision": 0.005990341232345204, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0003778566735238687}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_recall": 0.007150893881572016, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0005954055858899312}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_fmeasure": 0.005443826668587504, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0003144404879778184}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_precision": 0.09176826957181027, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0012268770230593445}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_recall": 0.09444076989884408, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.0014215142609522416}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_fmeasure": 0.08046273244596494, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0009128776785211843}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_precision": 0.11011996762454923, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0014803727480206098}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_recall": 0.11312057420513341, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.0016882892537450165}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_fmeasure": 0.09737556397488158, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0011716121889168436}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "bleu": 0.42527357697495255, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.06953287436188474}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 2, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-wiki_lingua_en_rephrase_en_3.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_precision": 0.10053229311714834, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0018186300913397432}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_recall": 0.10555528455488752, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.0020869634882788126}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_fmeasure": 0.08745924967780058, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.00145734596535055}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_precision": 0.007244189708893669, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.00047064519532168336}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_recall": 0.009481343391595305, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0007029880713328167}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_fmeasure": 0.0066746059699027156, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.000406229568602048}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_precision": 0.08113908472999296, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0014422772176468318}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_recall": 0.0854911553393767, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.001688161117709379}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_fmeasure": 0.06994035855230117, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0011004649845343271}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_precision": 0.09576776212352223, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0017136097429588115}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_recall": 0.10032211171110292, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.001963048810667937}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_fmeasure": 0.08314051391123733, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.001363454598942612}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "bleu": 0.5943947138668615, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.04701176872890474}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 3, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-wiki_lingua_en_rephrase_en_4.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_precision": 0.035781196323287794, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0015508373183201957}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_recall": 0.03889017824317372, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.0017168088677622815}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_fmeasure": 0.03034948263830278, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0012288540754404823}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_precision": 0.00393091834595338, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0004260496428881099}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_recall": 0.0052037277375677786, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.000540702114853635}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_fmeasure": 0.003519391416781204, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0003344402761329188}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_precision": 0.02946476725548691, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0012750099020740124}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_recall": 0.03200016757522908, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.0013894967755151602}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_fmeasure": 0.02470528087251438, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0009707901157133207}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_precision": 0.03393706130302841, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0014660077905268025}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_recall": 0.03664599115059629, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.0016081833996057495}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_fmeasure": 0.028653137000749736, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0011499184993932895}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "bleu": 0.06155277062691283, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.008573159858072549}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 4, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-wiki_lingua_en_rephrase_en_5.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_precision": 0.005466518578601791, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0006360451019998415}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_recall": 0.005849281181457089, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.00074547507260093}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_fmeasure": 0.004552330658915172, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0005166062646006778}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_precision": 0.0007290381746496603, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.00017635675582582084}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_recall": 0.0008879185087143063, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0002685233025455508}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_fmeasure": 0.0005748508715984116, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0001533404190150482}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_precision": 0.004641407616449875, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0005531411811140234}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_recall": 0.004667322692384839, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.0005697941762332262}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_fmeasure": 0.0037163176139750516, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0004064756060554951}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_precision": 0.005301656522138449, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0006174490006150588}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_recall": 0.005552151354486044, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.0006921370486199647}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_fmeasure": 0.00434986186087182, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.00048298225975487925}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "bleu": 9.384428445264235e-13, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "753f0a46-aeff-4cd2-932c-8548897cebe5", "prompt_jinja": "{{source}}\n\nHow would you rephrase that briefly in English? ||| {{target}}", "prompt_original_task": true, "comment": "", "bleu_stderr": 2.048496076777406e-11}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 5, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-wiki_lingua_en_summarize_above_en_0.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_precision": 0.06083058515622332, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0015146719263047607}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_recall": 0.08130079272174061, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.0017565791128464651}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_fmeasure": 0.06398795450213274, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0014079644213715252}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_precision": 0.0050127471219516785, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.00033788530642644006}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_recall": 0.006932802618765814, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.00047568921018848896}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_fmeasure": 0.005239039550783603, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0003252538524207895}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_precision": 0.056280510609653965, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0013608121285239516}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_recall": 0.07622653073256713, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.0016225806985423702}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_fmeasure": 0.05958527780779474, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0012805893801553889}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_precision": 0.05661921621032954, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.001398882606899158}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_recall": 0.0762057149480866, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.00164425094307365}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_fmeasure": 0.05971017630902041, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0013030395843137727}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "bleu": 0.2622610605101529, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.040142958508706195}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 0, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-wiki_lingua_en_summarize_above_en_1.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_precision": 0.11326508510637766, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0015009170735602506}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_recall": 0.1085356303020659, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.0014115882108590343}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_fmeasure": 0.09793700832399474, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0011273790641300194}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_precision": 0.004465655727904388, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.00028238368212925503}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_recall": 0.004056065876580251, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0002691375832689169}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_fmeasure": 0.0037395950495648657, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.00022671756481262202}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_precision": 0.09142917649433346, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0011841565139209087}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_recall": 0.0879166314153642, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.0011318568380660808}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_fmeasure": 0.07862622799687088, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0008488937771314072}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_precision": 0.1093895229104316, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0014332473740733347}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_recall": 0.10513740023129456, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.0013659288669536467}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_fmeasure": 0.09470526990675743, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0010805621439727475}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "bleu": 0.2382502920735501, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.02364732083952621}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 1, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-wiki_lingua_en_summarize_above_en_2.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_precision": 0.1195188168282025, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.001703506896411307}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_recall": 0.1202411166531759, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.0016720477170315902}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_fmeasure": 0.10460998387087274, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0012689330036383434}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_precision": 0.007701756834163109, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0004826378231441523}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_recall": 0.007828279383704012, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.00048735370780030713}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_fmeasure": 0.006572282621218325, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.00036858964749543314}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_precision": 0.09584725807304968, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0013390205513148664}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_recall": 0.09711441691325796, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.0013255057315948026}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_fmeasure": 0.08348917343043502, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0009381191407596309}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_precision": 0.11507270211895362, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0016248601270300064}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_recall": 0.11590379555656585, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.0015972136190990915}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_fmeasure": 0.10076143309185726, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0012074937805280613}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "bleu": 0.590846188853053, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.08622026318742701}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 2, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-wiki_lingua_en_summarize_above_en_3.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_precision": 0.10561107711679257, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.002135116442535178}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_recall": 0.10039756771582228, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.0018998579014479846}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_fmeasure": 0.08635223714862443, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.001447627888207993}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_precision": 0.010532137196295798, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0008099525943800139}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_recall": 0.009664993083625497, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0006422735345150598}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_fmeasure": 0.00773667698720356, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.00044084771240741574}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_precision": 0.08698250921011949, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0017942983133012848}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_recall": 0.08314557123023888, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.0015976310654642362}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_fmeasure": 0.07053215664409358, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0011503059093946877}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_precision": 0.10076653862384102, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.002033643570526417}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_recall": 0.09561293861925456, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.001792928927851378}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_fmeasure": 0.08220026138629714, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0013594564836760387}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "bleu": 0.5908356190981445, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.04489447675590248}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 3, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-wiki_lingua_en_summarize_above_en_4.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_precision": 0.035705986442768456, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.001618640602215546}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_recall": 0.03197561607719162, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.0014717555351632275}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_fmeasure": 0.027402695270076872, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.001157724691879728}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_precision": 0.00401479182015976, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.00043777035338946497}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_recall": 0.003920400654913368, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0004172562730537138}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_fmeasure": 0.00317018932471084, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.00031714568787250826}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_precision": 0.03013013623521446, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0013844671087526951}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_recall": 0.026445775089489882, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.0011917063204159026}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_fmeasure": 0.022642473398844034, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0009329018940160789}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_precision": 0.033971789988392904, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.001538890119076818}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_recall": 0.03041146272732838, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.0013938550134374181}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_fmeasure": 0.025956946800132486, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0010846296784172967}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "bleu": 0.029093100164270518, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.005798545611396172}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 4, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-wiki_lingua_en_summarize_above_en_5.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_precision": 0.00595043307989505, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.000767651406351477}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_recall": 0.004778481595486118, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.0006233197620300213}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_fmeasure": 0.00450639531355031, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.000555584129480168}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_precision": 0.0012396636176066488, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.00034778633785854403}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_recall": 0.0008030145717563384, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.00020935765891227752}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_fmeasure": 0.0007838553827958208, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.00021051541322103913}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_precision": 0.005169809582323384, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.0006759018705997662}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_recall": 0.004164169546014226, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.0005467863316598697}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_fmeasure": 0.0039016663853958123, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.000485959181689575}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_precision": 0.005734093038513017, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0007492388147769349}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_recall": 0.004562529353283875, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.00059727108953185}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_fmeasure": 0.004298574460690419, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0005305976692789258}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "bleu": 2.233297948144298e-16, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "088288f3-7516-4cf7-9406-0e082053bf54", "prompt_jinja": "{{source}}\n\n===\n\nWrite a summary of the text above in English : ||| {{target}}", "prompt_original_task": true, "comment": "", "bleu_stderr": 7.375838022248224e-15}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 5, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-wiki_lingua_en_tldr_en_0.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rouge1_precision": 0.10367122573898108, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "d3c5baa3-5e37-46f8-b1b2-5b834181c9da", "prompt_jinja": "{{source}}\n\nTL;DR in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0016311453272566847}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rouge1_recall": 0.1621848426933469, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "d3c5baa3-5e37-46f8-b1b2-5b834181c9da", "prompt_jinja": "{{source}}\n\nTL;DR in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.002239905291843337}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rouge1_fmeasure": 0.11655773215808052, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "d3c5baa3-5e37-46f8-b1b2-5b834181c9da", "prompt_jinja": "{{source}}\n\nTL;DR in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0015844131287312152}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rouge2_precision": 0.012276074546298633, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "d3c5baa3-5e37-46f8-b1b2-5b834181c9da", "prompt_jinja": "{{source}}\n\nTL;DR in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0005003323251677242}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rouge2_recall": 0.020862324606727274, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "d3c5baa3-5e37-46f8-b1b2-5b834181c9da", "prompt_jinja": "{{source}}\n\nTL;DR in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0009624161724708909}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rouge2_fmeasure": 0.014011218070534905, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "d3c5baa3-5e37-46f8-b1b2-5b834181c9da", "prompt_jinja": "{{source}}\n\nTL;DR in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0005398251030136181}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rougeL_precision": 0.08966835876675931, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "d3c5baa3-5e37-46f8-b1b2-5b834181c9da", "prompt_jinja": "{{source}}\n\nTL;DR in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.00130510256531558}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rougeL_recall": 0.14342974130461322, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "d3c5baa3-5e37-46f8-b1b2-5b834181c9da", "prompt_jinja": "{{source}}\n\nTL;DR in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.0019375952453362063}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rougeL_fmeasure": 0.10161863320310922, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "d3c5baa3-5e37-46f8-b1b2-5b834181c9da", "prompt_jinja": "{{source}}\n\nTL;DR in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0012820750167481308}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rougeLsum_precision": 0.09728691738451277, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "d3c5baa3-5e37-46f8-b1b2-5b834181c9da", "prompt_jinja": "{{source}}\n\nTL;DR in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.0015087459211237015}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rougeLsum_recall": 0.1530554381860451, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "d3c5baa3-5e37-46f8-b1b2-5b834181c9da", "prompt_jinja": "{{source}}\n\nTL;DR in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.0021142656537031777}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rougeLsum_fmeasure": 0.10956185275309588, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "d3c5baa3-5e37-46f8-b1b2-5b834181c9da", "prompt_jinja": "{{source}}\n\nTL;DR in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0014665536918047877}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "bleu": 0.7280055452213583, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "d3c5baa3-5e37-46f8-b1b2-5b834181c9da", "prompt_jinja": "{{source}}\n\nTL;DR in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.06972461776949417}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 0, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}
146m14b14b/eval/agg.lm1-146m-14b_GEM-wiki_lingua_en_tldr_en_1.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"results": [{"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rouge1_precision": 0.11922048835394526, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "d3c5baa3-5e37-46f8-b1b2-5b834181c9da", "prompt_jinja": "{{source}}\n\nTL;DR in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_precision_stderr": 0.0017010312663949007}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rouge1_recall": 0.12373373738142388, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "d3c5baa3-5e37-46f8-b1b2-5b834181c9da", "prompt_jinja": "{{source}}\n\nTL;DR in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_recall_stderr": 0.0018173643062502098}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rouge1_fmeasure": 0.10548837365464835, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "d3c5baa3-5e37-46f8-b1b2-5b834181c9da", "prompt_jinja": "{{source}}\n\nTL;DR in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge1_fmeasure_stderr": 0.0012991497483475392}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rouge2_precision": 0.008234339790533517, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "d3c5baa3-5e37-46f8-b1b2-5b834181c9da", "prompt_jinja": "{{source}}\n\nTL;DR in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_precision_stderr": 0.0005690033754874188}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rouge2_recall": 0.009356412797919933, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "d3c5baa3-5e37-46f8-b1b2-5b834181c9da", "prompt_jinja": "{{source}}\n\nTL;DR in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_recall_stderr": 0.0005938642159151799}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rouge2_fmeasure": 0.007230676056378635, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "d3c5baa3-5e37-46f8-b1b2-5b834181c9da", "prompt_jinja": "{{source}}\n\nTL;DR in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rouge2_fmeasure_stderr": 0.0004084061340426134}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rougeL_precision": 0.09596744652257369, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "d3c5baa3-5e37-46f8-b1b2-5b834181c9da", "prompt_jinja": "{{source}}\n\nTL;DR in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_precision_stderr": 0.001368600668118387}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rougeL_recall": 0.09975162242794412, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "d3c5baa3-5e37-46f8-b1b2-5b834181c9da", "prompt_jinja": "{{source}}\n\nTL;DR in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_recall_stderr": 0.0014345772942689061}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rougeL_fmeasure": 0.0843032809934075, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "d3c5baa3-5e37-46f8-b1b2-5b834181c9da", "prompt_jinja": "{{source}}\n\nTL;DR in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeL_fmeasure_stderr": 0.0009705899345773655}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rougeLsum_precision": 0.11440643074566167, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "d3c5baa3-5e37-46f8-b1b2-5b834181c9da", "prompt_jinja": "{{source}}\n\nTL;DR in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_precision_stderr": 0.001619972437724354}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rougeLsum_recall": 0.11862725022797178, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "d3c5baa3-5e37-46f8-b1b2-5b834181c9da", "prompt_jinja": "{{source}}\n\nTL;DR in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_recall_stderr": 0.0017155423340356613}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rougeLsum_fmeasure": 0.10113335490345467, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "d3c5baa3-5e37-46f8-b1b2-5b834181c9da", "prompt_jinja": "{{source}}\n\nTL;DR in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "rougeLsum_fmeasure_stderr": 0.0012233443711385755}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "bleu": 0.5504071981349774, "fixed_answer_choice_list": null, "dataset_path": "GEM/wiki_lingua", "dataset_name": "en", "subset": null, "prompt_id": "d3c5baa3-5e37-46f8-b1b2-5b834181c9da", "prompt_jinja": "{{source}}\n\nTL;DR in English: ||| {{target}}", "prompt_original_task": true, "comment": "", "bleu_stderr": 0.0561562373694907}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-146m-14b/transformers,use_accelerate=True,tokenizer=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/gpt2,dtype=bfloat16", "task_args": "", "num_fewshot": 1, "batch_size": 16, "device": "cuda", "use_cache": false, "limit": 3000, "bootstrap_iters": 10, "seed": 1234}}