Hugging Face's logo

datablations
/
lm1-misc-pile

TensorBoard
Model card Files Metrics Community
Muennighoff commited on
Commit
c305798
•
1 Parent(s): 46fa37e

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. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-web_nlg_en_PALM_prompt_0.json +1 -0
  2. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-web_nlg_en_PALM_prompt_1.json +1 -0
  3. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-web_nlg_en_PALM_prompt_2.json +1 -0
  4. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-web_nlg_en_PALM_prompt_3.json +1 -0
  5. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-web_nlg_en_PALM_prompt_4.json +1 -0
  6. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-web_nlg_en_PALM_prompt_5.json +1 -0
  7. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-web_nlg_en_explicit-graph-description2_0.json +1 -0
  8. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-web_nlg_en_explicit-graph-description2_1.json +1 -0
  9. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-web_nlg_en_explicit-graph-description2_2.json +1 -0
  10. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-web_nlg_en_explicit-graph-description2_3.json +1 -0
  11. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-web_nlg_en_explicit-graph-description2_4.json +1 -0
  12. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-web_nlg_en_explicit-graph-description2_5.json +1 -0
  13. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-web_nlg_en_implicit-graph-description_0.json +1 -0
  14. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-web_nlg_en_implicit-graph-description_1.json +1 -0
  15. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-web_nlg_en_implicit-graph-description_2.json +1 -0
  16. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-web_nlg_en_implicit-graph-description_3.json +1 -0
  17. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-web_nlg_en_implicit-graph-description_4.json +1 -0
  18. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-web_nlg_en_implicit-graph-description_5.json +1 -0
  19. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-web_nlg_en_non-explicit-description_0.json +1 -0
  20. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-web_nlg_en_non-explicit-description_1.json +1 -0
  21. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-web_nlg_en_non-explicit-description_2.json +1 -0
  22. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-web_nlg_en_non-explicit-description_3.json +1 -0
  23. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-web_nlg_en_non-explicit-description_4.json +1 -0
  24. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-web_nlg_en_non-explicit-description_5.json +1 -0
  25. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-web_nlg_en_very-explicit-description_0.json +1 -0
  26. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-web_nlg_en_very-explicit-description_1.json +1 -0
  27. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-web_nlg_en_very-explicit-description_2.json +1 -0
  28. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-web_nlg_en_very-explicit-description_3.json +1 -0
  29. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-web_nlg_en_very-explicit-description_4.json +1 -0
  30. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-web_nlg_en_very-explicit-description_5.json +1 -0
  31. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-wiki_lingua_en_article_summary_en_0.json +1 -0
  32. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-wiki_lingua_en_article_summary_en_1.json +1 -0
  33. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-wiki_lingua_en_article_summary_en_2.json +1 -0
  34. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-wiki_lingua_en_article_summary_en_3.json +1 -0
  35. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-wiki_lingua_en_article_summary_en_4.json +1 -0
  36. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-wiki_lingua_en_article_summary_en_5.json +1 -0
  37. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-wiki_lingua_en_rephrase_en_0.json +1 -0
  38. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-wiki_lingua_en_rephrase_en_1.json +1 -0
  39. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-wiki_lingua_en_rephrase_en_2.json +1 -0
  40. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-wiki_lingua_en_rephrase_en_3.json +1 -0
  41. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-wiki_lingua_en_rephrase_en_4.json +1 -0
  42. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-wiki_lingua_en_rephrase_en_5.json +1 -0
  43. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-wiki_lingua_en_summarize_above_en_0.json +1 -0
  44. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-wiki_lingua_en_summarize_above_en_1.json +1 -0
  45. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-wiki_lingua_en_summarize_above_en_2.json +1 -0
  46. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-wiki_lingua_en_summarize_above_en_3.json +1 -0
  47. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-wiki_lingua_en_summarize_above_en_4.json +1 -0
  48. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-wiki_lingua_en_summarize_above_en_5.json +1 -0
  49. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-wiki_lingua_en_tldr_en_0.json +1 -0
  50. 220m7b57b5/eval/agg.lm1-220m-7b5_GEM-wiki_lingua_en_tldr_en_1.json +1 -0
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.2470132389281794, "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": 0.024002041972246133}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_precision": 0.05869608248520842, "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": 0.0023418437832201868}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_recall": 0.20816537313165334, "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": 0.004363240809955283}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_fmeasure": 0.07945556760491934, "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": 0.0018763238861836877}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_precision": 0.025554038423215544, "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": 0.0011939821439251124}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_recall": 0.10140329387715467, "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": 0.0028057725221001355}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_fmeasure": 0.03660833438150841, "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": 0.001100777714916583}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_precision": 0.05722355323316198, "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": 0.0022447001844704303}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_recall": 0.2053626360936232, "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": 0.0042990838339060505}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_fmeasure": 0.07779520657098479, "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": 0.0017915988634161517}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_precision": 0.05665115836870732, "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": 0.0022509535075962666}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_recall": 0.20167181058435762, "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": 0.004114520880555949}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_fmeasure": 0.07670579662288195, "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": 0.0017734296978415465}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.2855476187137642, "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": 0.03227733703551035}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_precision": 0.06524544825277895, "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": 0.0021779484868505146}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_recall": 0.21139563660239885, "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": 0.003964920300089838}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_fmeasure": 0.08765527786557152, "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": 0.0019897193030688743}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_precision": 0.025767761662064487, "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": 0.0011672938952980172}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_recall": 0.09430977366190377, "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": 0.0025953062420921275}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_fmeasure": 0.036434969681400774, "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": 0.001155467893994337}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_precision": 0.06156492223032873, "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": 0.0020523056730015823}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_recall": 0.2038680112360577, "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": 0.003812943348220746}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_fmeasure": 0.08304491621550703, "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": 0.0018059904505394503}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_precision": 0.06197763706322762, "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": 0.0020794221174581487}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_recall": 0.20355852337211516, "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": 0.003764546042324205}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_fmeasure": 0.08339837790352603, "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": 0.0018329339066338202}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.2864274010162566, "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": 0.02830546364175333}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_precision": 0.06783215358771556, "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": 0.002090888747972165}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_recall": 0.22138587795535367, "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": 0.003903037202789631}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_fmeasure": 0.0925124710513938, "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": 0.0020869881906242136}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_precision": 0.026875228585622548, "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": 0.001070936525488954}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_recall": 0.0988492806414047, "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": 0.002610711343589784}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_fmeasure": 0.038262581855489086, "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": 0.0011712844078541956}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_precision": 0.06297992735051472, "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": 0.0018483585802301025}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_recall": 0.2134940852293742, "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": 0.003778189492592616}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_fmeasure": 0.08697513104280281, "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": 0.0018518665685962159}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_precision": 0.0635465636926062, "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": 0.001900104766598554}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_recall": 0.21323474003372503, "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": 0.003751798074873092}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_fmeasure": 0.08736568328195454, "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": 0.001888088986617876}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.2864048655486057, "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": 0.037454971776838}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_precision": 0.06740222556506215, "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": 0.002210504181928137}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_recall": 0.22956928475213337, "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": 0.003997241926123326}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_fmeasure": 0.0912440310361926, "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": 0.002057320180020397}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_precision": 0.026876421531834702, "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": 0.0011206870736437747}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_recall": 0.10214742165037678, "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": 0.002646201869660272}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_fmeasure": 0.038237236474277785, "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": 0.001183860044481188}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_precision": 0.06307609514566138, "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": 0.0020021872586120397}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_recall": 0.22156818068672537, "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": 0.0038500842834299346}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_fmeasure": 0.08625410231524447, "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": 0.0018422312969977972}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_precision": 0.06344776968110694, "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": 0.002026499548496223}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_recall": 0.2215545857990353, "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": 0.0038440696387562643}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_fmeasure": 0.08655326887562105, "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": 0.0018638494278753182}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.30299260397701133, "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": 0.03776599386559902}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_precision": 0.07140612607619035, "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": 0.0022021853930916176}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_recall": 0.23914100173824104, "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": 0.003935999196689264}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_fmeasure": 0.09739252402210666, "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": 0.002168582806484753}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_precision": 0.031023668868501216, "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": 0.0013538425626279963}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_recall": 0.10947572550063607, "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": 0.002747589012727159}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_fmeasure": 0.042444264913734785, "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": 0.0013172139194447315}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_precision": 0.06583563215149338, "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": 0.001863780080394457}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_recall": 0.2299802944684882, "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": 0.003766169713586082}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_fmeasure": 0.09141250194289818, "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": 0.0019071063232336046}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_precision": 0.06707161047873238, "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": 0.0019840908441180363}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_recall": 0.23027746607274016, "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": 0.003738900975653066}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_fmeasure": 0.09224313841851178, "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": 0.0019562908112556432}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.29766196455460725, "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": 0.03056279900060948}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_precision": 0.07457191719751274, "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": 0.0024562285116543296}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_recall": 0.24169660652155414, "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": 0.003946464593048434}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge1_fmeasure": 0.09919130134598364, "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": 0.0022813442744688485}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_precision": 0.032773996107936534, "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": 0.0015839604927304}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_recall": 0.1095327777694728, "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": 0.0027384254788386984}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rouge2_fmeasure": 0.04315943772956645, "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": 0.0014448380001542028}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_precision": 0.0691396900165197, "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": 0.0021201246483540464}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_recall": 0.2329967123384895, "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": 0.0037887799368644215}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeL_fmeasure": 0.09336945041280252, "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": 0.002013899382003858}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_precision": 0.07003694615232588, "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": 0.0022144247071161968}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_recall": 0.23283936287872428, "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": 0.0037634958657852477}, {"task_name": "GEM/web_nlg_en", "prompt_name": "PALM_prompt", "rougeLsum_fmeasure": 0.09391276057220475, "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": 0.002052779259520063}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.0248617903991397, "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": 0.0011434850560320853}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_recall": 0.1411098785046503, "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": 0.001957418108907302}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_fmeasure": 0.03682449474664423, "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": 0.001075644895354243}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_precision": 0.0007716668827793864, "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": 0.00021269711797048112}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_recall": 0.0019537719790670887, "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": 0.0003582685451874752}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_fmeasure": 0.000842345548175766, "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": 0.0002153717552554103}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_precision": 0.023331890995755642, "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": 0.0009146032826694152}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_recall": 0.1392558839352905, "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": 0.00188398935218735}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_fmeasure": 0.03537063504919988, "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": 0.0008837014987387857}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_precision": 0.016566330901198578, "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": 0.0009109548374086303}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_recall": 0.09380553249433002, "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": 0.001482408777774195}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_fmeasure": 0.023889371081173, "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": 0.000847324247546207}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "bleu": 0.003304656924332801, "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": 0.0002357303417870464}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.17193868199001666, "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": 0.002526664633506342}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_recall": 0.16092252608921864, "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": 0.0025363721875677125}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_fmeasure": 0.14557616214341593, "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": 0.0018082502726000061}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_precision": 0.017822604334142886, "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": 0.001163098557262362}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_recall": 0.017824877216273783, "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": 0.0010953948712234692}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_fmeasure": 0.01478573154051706, "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": 0.0008419045118869184}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_precision": 0.14455425222028767, "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": 0.002152971433485278}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_recall": 0.1351485919438946, "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": 0.002121399411940993}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_fmeasure": 0.12127429715169136, "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": 0.0014306811162337182}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_precision": 0.1531028156588724, "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": 0.002279823151167151}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_recall": 0.14235486646030968, "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": 0.002191433538866353}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_fmeasure": 0.12865712218171224, "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": 0.001556095888289526}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "bleu": 0.7684277465132661, "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": 0.1333707947952387}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.22882473247907112, "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": 0.0038518788451608926}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_recall": 0.29902689022333023, "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": 0.0049105220426749205}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_fmeasure": 0.21273774790890274, "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": 0.002862298904811987}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_precision": 0.06505529355658353, "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": 0.002554842255013279}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_recall": 0.09595036644512045, "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": 0.0030736316716887158}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_fmeasure": 0.061302291693352925, "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": 0.0020239566861584026}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_precision": 0.1934489018681778, "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": 0.003387814428193738}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_recall": 0.2575610697212871, "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": 0.004406926115400504}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_fmeasure": 0.1792983798884188, "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": 0.00241876765302595}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_precision": 0.20334040984836957, "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": 0.003511792550501819}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_recall": 0.2630663806902787, "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": 0.004251688329429685}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_fmeasure": 0.18748250731872323, "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": 0.0025282235134524876}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "bleu": 1.0144425667807024, "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": 0.10905315088614692}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.2730815312972432, "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": 0.004825231533655922}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_recall": 0.33489751702159615, "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": 0.005019362675201049}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_fmeasure": 0.24405575384062297, "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": 0.003390620639096825}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_precision": 0.09147579976257764, "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": 0.003142239701007767}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_recall": 0.12181501911692545, "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": 0.003281460863415726}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_fmeasure": 0.08294640662873075, "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": 0.002466976544448955}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_precision": 0.23009219757904717, "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": 0.004160956624281562}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_recall": 0.2893843077489737, "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": 0.0045371766484929126}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_fmeasure": 0.2056899991458418, "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": 0.0028738202404970086}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_precision": 0.2408276500835719, "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": 0.004336511332873335}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_recall": 0.2921676768826999, "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": 0.004310890720533993}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_fmeasure": 0.21376985343130714, "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": 0.002991125558443766}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "bleu": 1.324337486227112, "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": 0.08529407489265287}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.3004668922387791, "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": 0.005202327993278515}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_recall": 0.3381015117634111, "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": 0.004926181256404199}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_fmeasure": 0.2573966620241667, "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": 0.00360782853105973}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_precision": 0.1135229703501693, "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": 0.0035770821938528497}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_recall": 0.13090277247230048, "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": 0.00334170274274197}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_fmeasure": 0.09605879480553722, "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": 0.0026897626761239342}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_precision": 0.2579408245927597, "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": 0.004619853165223231}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_recall": 0.2949912956027357, "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": 0.004439834001883179}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_fmeasure": 0.22027925029754633, "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": 0.00309656820864388}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_precision": 0.2674141566743051, "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": 0.004769770402583653}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_recall": 0.2965330042570855, "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": 0.004256557776307167}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_fmeasure": 0.2270423544148196, "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": 0.0032084127495861873}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "bleu": 1.4371640698778811, "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": 0.10400308604922297}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.31389086214595796, "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": 0.005511590700185723}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_recall": 0.3432479229856668, "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": 0.004780112256990474}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge1_fmeasure": 0.26457662822391115, "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": 0.0037120067585936093}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_precision": 0.128182260271063, "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": 0.004048216640966436}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_recall": 0.1375596549218887, "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": 0.003287107225763228}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rouge2_fmeasure": 0.10380910940230743, "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": 0.00279856847849847}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_precision": 0.2703532594581765, "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": 0.004903667924990887}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_recall": 0.2997975976677086, "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": 0.004274831930821855}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeL_fmeasure": 0.22714153548731583, "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": 0.003197798431801004}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_precision": 0.2811052155144168, "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": 0.005067409241664891}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_recall": 0.3035198712824658, "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": 0.004162619127613586}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "rougeLsum_fmeasure": 0.23490404484580202, "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": 0.0033092693657403474}, {"task_name": "GEM/web_nlg_en", "prompt_name": "explicit-graph-description2", "bleu": 1.426813356520762, "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": 0.07637434838106619}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.08334994279019105, "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": 0.018748309082212023}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_precision": 0.030070455540385104, "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": 0.0007549842191224862}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_recall": 0.23192020207522754, "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": 0.0027859327621759435}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_fmeasure": 0.04988655480401325, "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": 0.0008236285402288305}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_precision": 0.0038552559833399595, "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": 0.00024159123811971067}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_recall": 0.03372092008925444, "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": 0.0017150196698228661}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_fmeasure": 0.006620886187941369, "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": 0.00039462015364294286}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_precision": 0.029396226905775355, "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": 0.0007073526772596732}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_recall": 0.2289002484039579, "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": 0.002748861828734789}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_fmeasure": 0.04893516033464724, "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": 0.0007880409616162576}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_precision": 0.022554350647870147, "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": 0.0006997403740778779}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_recall": 0.17679509299725407, "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": 0.0023967420151251085}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_fmeasure": 0.03706372138195696, "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": 0.0006945085675693861}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.7257952302819668, "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": 0.1273069433851748}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_precision": 0.1657234988993074, "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": 0.002567846275536951}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_recall": 0.16021335289487582, "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": 0.002740661059769588}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_fmeasure": 0.14167529394488848, "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": 0.0018924466602806278}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_precision": 0.017981060701813652, "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": 0.0011649739165457932}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_recall": 0.01969011125673082, "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": 0.0012089300876458326}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_fmeasure": 0.01535561520309398, "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": 0.0008727895396409643}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_precision": 0.13933426569455393, "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": 0.0021889015963596136}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_recall": 0.13498672608104417, "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": 0.0023604246095271448}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_fmeasure": 0.11807186468862133, "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": 0.0015144892688643149}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_precision": 0.14736164369432678, "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": 0.002316919994284117}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_recall": 0.14146144963746662, "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": 0.0023621122021278526}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_fmeasure": 0.12504823161412182, "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": 0.00163478797997344}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.8846555651917197, "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": 0.08142013626852958}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_precision": 0.2077035120584479, "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": 0.003800118155832393}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_recall": 0.31966138271975814, "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": 0.005119077460794474}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_fmeasure": 0.19929270987633105, "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": 0.0027385624944381037}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_precision": 0.059668061023018824, "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": 0.0023660417929754545}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_recall": 0.11236661625797283, "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": 0.0032800608381254705}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_fmeasure": 0.0597704716961618, "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": 0.0018634494046912602}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_precision": 0.17713318087352298, "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": 0.003315487340297903}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_recall": 0.28382124839738937, "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": 0.004850911382885647}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_fmeasure": 0.17060240499732354, "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": 0.0023361549717732335}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_precision": 0.18407367153416493, "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": 0.0034779619153249815}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_recall": 0.27775546966623693, "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": 0.004357079972672227}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_fmeasure": 0.1749056810880912, "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": 0.0024557066632072203}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 1.0398538260324774, "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": 0.07824794254019525}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_precision": 0.2389983247359521, "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": 0.004511729074401231}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_recall": 0.3587416075240466, "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": 0.005046170970680584}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_fmeasure": 0.22416300826387658, "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": 0.0030921508501741417}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_precision": 0.08136848635282187, "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": 0.0027581931272128903}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_recall": 0.1413307908669179, "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": 0.0034695197162447383}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_fmeasure": 0.0790117294260896, "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": 0.002180141487152526}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_precision": 0.203320192274512, "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": 0.003852417463967525}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_recall": 0.31980048584652565, "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": 0.004825602773280325}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_fmeasure": 0.19222474838483417, "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": 0.0026240648276200066}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_precision": 0.2108854696562002, "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": 0.004071014032359802}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_recall": 0.31215644006725296, "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": 0.004333979448826886}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_fmeasure": 0.19662916172177014, "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": 0.0027710185945962677}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.9928640647621334, "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": 0.07630427189860162}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_precision": 0.2470229974332293, "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": 0.004730824733275526}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_recall": 0.3666911864611959, "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": 0.005054207652920528}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_fmeasure": 0.2271992309006609, "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": 0.0033064521653597616}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_precision": 0.0889055889558497, "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": 0.0029033662382457843}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_recall": 0.14963481592169695, "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": 0.0035479909748617443}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_fmeasure": 0.08406534333433156, "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": 0.00234196493550534}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_precision": 0.21312177044434366, "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": 0.004187406074549313}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_recall": 0.33003769745963074, "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": 0.00486007718156714}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_fmeasure": 0.1970169482477049, "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": 0.0028742175496318982}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_precision": 0.2193683377976872, "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": 0.0043711369660711505}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_recall": 0.32011281291177024, "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": 0.004351847501804956}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_fmeasure": 0.19983771726616284, "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": 0.0030074393097768073}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 1.0853105256942843, "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": 0.08756095828849798}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_precision": 0.2577193492576544, "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": 0.004961882283280986}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_recall": 0.3787841404647953, "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": 0.0049147710594367915}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge1_fmeasure": 0.2376138291487847, "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": 0.003296870506654906}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_precision": 0.099464377753478, "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": 0.0031870297946555532}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_recall": 0.15661648161646427, "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": 0.0034506929154406334}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rouge2_fmeasure": 0.09142528546583407, "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": 0.0023426059953661534}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_precision": 0.2222245289227495, "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": 0.0043305271968674015}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_recall": 0.3406223745481442, "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": 0.004671722829085883}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeL_fmeasure": 0.20685989535839744, "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": 0.0028654031656975237}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_precision": 0.22906137819697442, "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": 0.004549678767902717}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_recall": 0.33208650809863727, "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": 0.004238667525429603}, {"task_name": "GEM/web_nlg_en", "prompt_name": "implicit-graph-description", "rougeLsum_fmeasure": 0.2100644010616573, "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": 0.003007725178547181}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.04638900916509621, "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": 0.001381574654990731}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_recall": 0.18225688622589034, "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": 0.0025564819722441374}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_fmeasure": 0.05785562404770463, "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": 0.0010580886851143255}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_precision": 0.0031049905665538518, "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": 0.0002294877760634393}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_recall": 0.020249681594683613, "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": 0.001266890639043393}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_fmeasure": 0.004839284419441053, "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": 0.0003142723165291525}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_precision": 0.04307034272835799, "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": 0.0012049508438658615}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_recall": 0.17613780593439127, "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": 0.002513632912368753}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_fmeasure": 0.05471027256014058, "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": 0.0009541568628228309}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_precision": 0.04116906029494233, "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": 0.0012839464709169964}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_recall": 0.15781011704062736, "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": 0.002137044269825738}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_fmeasure": 0.05041002107948325, "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": 0.0009467047128866438}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "bleu": 0.050711652769452574, "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": 0.011214962522358278}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.1693826680244151, "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": 0.0027547636280590196}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_recall": 0.21063133932730638, "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": 0.0040607584716417415}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_fmeasure": 0.1553880799162404, "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": 0.0021412107596476713}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_precision": 0.026640932753563176, "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": 0.0014964515466493716}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_recall": 0.04716115990415862, "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": 0.0022734949764752176}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_fmeasure": 0.02657894823260922, "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": 0.0012545355152779483}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_precision": 0.14243067697418593, "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": 0.0023457378304752863}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_recall": 0.1826038076256651, "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": 0.0037925378540135664}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_fmeasure": 0.1305486904847378, "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": 0.0017586404618870843}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_precision": 0.15002419508561124, "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": 0.0024941106497848716}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_recall": 0.18442412009786432, "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": 0.0034982664767672284}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_fmeasure": 0.1364296007053323, "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": 0.0018585747882451662}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "bleu": 0.9534893716298812, "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": 0.1295046344795696}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.20378699726362787, "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": 0.003499401908477442}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_recall": 0.29039673083630463, "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": 0.004888207311656959}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_fmeasure": 0.19131133812376055, "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": 0.002476526010152949}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_precision": 0.054634968223841666, "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": 0.002412169422109711}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_recall": 0.09726270833855312, "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": 0.003147645321841744}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_fmeasure": 0.05311952756303624, "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": 0.0017625237602222526}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_precision": 0.1743921312047424, "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": 0.00313469327664464}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_recall": 0.2582517841880744, "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": 0.004689854925384346}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_fmeasure": 0.16418426378825893, "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": 0.0021512732627924445}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_precision": 0.18144680257700982, "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": 0.003274259330271958}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_recall": 0.25506190053007244, "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": 0.004240940484642221}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_fmeasure": 0.16870970746087516, "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": 0.002241583575668234}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "bleu": 0.9742686315463768, "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": 0.06304865024693125}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.22927804158190548, "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": 0.0043077324327183576}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_recall": 0.3294081021915493, "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": 0.005148065368101898}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_fmeasure": 0.21282562230891222, "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": 0.003037449491429203}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_precision": 0.07825144347596119, "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": 0.003028440816234046}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_recall": 0.12801142438438318, "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": 0.0034748361764916813}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_fmeasure": 0.07357168720801384, "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": 0.0022418396556559157}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_precision": 0.19787725440544807, "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": 0.003803980393075029}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_recall": 0.29571856975050964, "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": 0.004929603754546493}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_fmeasure": 0.1842528650204582, "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": 0.0025981002359634258}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_precision": 0.20431722202402317, "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": 0.003974679118789089}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_recall": 0.28851412895390455, "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": 0.004452130914545327}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_fmeasure": 0.18751756928665156, "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": 0.0026992448187068067}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "bleu": 1.070241764666352, "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": 0.13372853017110806}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.2619800388575182, "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": 0.004783245683277002}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_recall": 0.34639413493795207, "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": 0.005002897920370203}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_fmeasure": 0.2356407835740911, "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": 0.0032675963226520134}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_precision": 0.10154033145219339, "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": 0.0033898899872025164}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_recall": 0.14446115226241912, "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": 0.0035610277886803346}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_fmeasure": 0.09142228086785963, "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": 0.002501625424189462}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_precision": 0.22704126812025263, "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": 0.004205577873021702}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_recall": 0.3115890114454386, "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": 0.004795026157826548}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_fmeasure": 0.20529916747778176, "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": 0.002824662241806233}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_precision": 0.2334029270612226, "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": 0.004374576696055797}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_recall": 0.3045160432395484, "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": 0.004344115062249235}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_fmeasure": 0.20813046842802627, "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": 0.0029195086928997006}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "bleu": 1.1141191654591824, "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": 0.07870221412087036}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.2780466507567164, "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": 0.00516214549353846}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_recall": 0.35082935541225824, "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": 0.00490713561072582}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge1_fmeasure": 0.2435734274073474, "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": 0.003429753227268956}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_precision": 0.1126294123523219, "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": 0.00373132040153298}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_recall": 0.14930398133068154, "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": 0.003477522073766097}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rouge2_fmeasure": 0.09723257871664684, "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": 0.0025431290052284053}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_precision": 0.2421462577916391, "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": 0.004606529598243439}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_recall": 0.3145972633270076, "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": 0.0046578650592379615}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeL_fmeasure": 0.2123142847759052, "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": 0.0029448367062216577}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_precision": 0.2501295208158327, "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": 0.004791408943905472}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_recall": 0.3094856112571341, "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": 0.004261043006602401}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "rougeLsum_fmeasure": 0.21654574972897772, "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": 0.003057778617604425}, {"task_name": "GEM/web_nlg_en", "prompt_name": "non-explicit-description", "bleu": 1.1076702355996768, "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": 0.04516609022262046}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.10877339345103203, "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": 0.0017676314022560438}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_recall": 0.5966297664887575, "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": 0.004357566028665667}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_fmeasure": 0.17448793075519228, "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": 0.0023788475299852327}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_precision": 0.040248602828145706, "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": 0.000974565397422352}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_recall": 0.23179382326447295, "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": 0.0034554710661058646}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_fmeasure": 0.06452901581107438, "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": 0.001372404833335715}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_precision": 0.08986765485052847, "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": 0.001232540970542953}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_recall": 0.533176214517191, "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": 0.004284493170471842}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_fmeasure": 0.1461562921920406, "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": 0.0016648189557296527}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_precision": 0.09064357871985745, "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": 0.0015859955091887365}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_recall": 0.49628166954580544, "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": 0.003871465689600495}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_fmeasure": 0.1451260252618243, "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": 0.002146032652253002}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "bleu": 0.7456292603754174, "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": 0.0834641122106676}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.11550360396098022, "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": 0.00247946160320685}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_recall": 0.3447812830852719, "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": 0.005068682828131058}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_fmeasure": 0.13957172000190018, "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": 0.0023563158981155363}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_precision": 0.029531462378906137, "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": 0.0013462575205621342}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_recall": 0.11223416303298638, "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": 0.0034752178341412414}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_fmeasure": 0.039371131493754236, "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": 0.0014605898807165082}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_precision": 0.10120492261222327, "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": 0.0020938999277697145}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_recall": 0.32221661384314343, "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": 0.004864037404384629}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_fmeasure": 0.12423853547056209, "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": 0.001973862248151174}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_precision": 0.10056006172307458, "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": 0.002264927518584629}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_recall": 0.2960283009580404, "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": 0.0044536210767821115}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_fmeasure": 0.12001147275059507, "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": 0.002099757151031158}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "bleu": 0.6252121265479059, "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": 0.08561481755044176}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.17978746094613066, "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": 0.003223269796033701}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_recall": 0.3991583853728954, "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": 0.005196252635965315}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_fmeasure": 0.19667965100080456, "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": 0.002403096154079486}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_precision": 0.05692160552689046, "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": 0.0020194812558206945}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_recall": 0.15843907751416444, "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": 0.003593218460427564}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_fmeasure": 0.0655421200388863, "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": 0.001655431015435781}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_precision": 0.15558060960116304, "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": 0.002760602832729305}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_recall": 0.3651396277557831, "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": 0.005094393120407086}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_fmeasure": 0.17205591725908315, "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": 0.0019955765272850604}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_precision": 0.15819057263024444, "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": 0.0029725373991650386}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_recall": 0.34684061489200496, "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": 0.004495598204481229}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_fmeasure": 0.17112834293465182, "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": 0.0021419093921029217}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "bleu": 0.8505803601192041, "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": 0.07686311815461677}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.19218576827837394, "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": 0.0038903326718722725}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_recall": 0.4241078446649215, "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": 0.004997755877660919}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_fmeasure": 0.20599763009776187, "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": 0.0026553485850209836}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_precision": 0.07354006358488266, "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": 0.0025647760852483647}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_recall": 0.18074074589214492, "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": 0.0035565122411799228}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_fmeasure": 0.07909682423009037, "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": 0.001844139139440938}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_precision": 0.1686879943141715, "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": 0.0034188355607413425}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_recall": 0.3915581412457528, "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": 0.004943563849410827}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_fmeasure": 0.18269679068630695, "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": 0.0022689544377994587}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_precision": 0.17084076805596005, "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": 0.003614281258074663}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_recall": 0.37018541561292334, "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": 0.004344462824974133}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_fmeasure": 0.18090166156905632, "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": 0.002412304850426574}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "bleu": 0.8232224220153507, "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": 0.07241444822510976}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.2145154894402684, "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": 0.0043504227534435095}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_recall": 0.4319311866342634, "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": 0.004841890049697201}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_fmeasure": 0.22094725474527804, "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": 0.00281918595948166}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_precision": 0.08967475350861655, "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": 0.0029253745282281658}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_recall": 0.192167033298445, "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": 0.003503137451514462}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_fmeasure": 0.09100243331728804, "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": 0.002008748453651486}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_precision": 0.18949865356185708, "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": 0.0038593402900291038}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_recall": 0.39932025651276526, "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": 0.004765127050124904}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_fmeasure": 0.19687286896005082, "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": 0.002417277695293234}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_precision": 0.1920606276236361, "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": 0.0040762331621287335}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_recall": 0.38005448554479504, "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": 0.004229479876011144}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_fmeasure": 0.19559622398761292, "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": 0.002588840248207576}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "bleu": 0.8832944903767832, "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": 0.057663141908889604}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.22205618914858594, "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": 0.004628884408025807}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_recall": 0.43490315424383064, "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": 0.004792522189978164}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge1_fmeasure": 0.22458403143720304, "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": 0.0029691363784075135}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_precision": 0.09607041419944147, "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": 0.0031305444273312076}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_recall": 0.19747036576528862, "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": 0.0035302794360299724}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rouge2_fmeasure": 0.0952366147047778, "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": 0.002097429198771996}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_precision": 0.19656903774141674, "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": 0.004120201165743963}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_recall": 0.4017037291667085, "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": 0.004715677239119104}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeL_fmeasure": 0.20023605221994797, "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": 0.0025423805596355526}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_precision": 0.1996263121342979, "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": 0.004349640784270575}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_recall": 0.3829286275310109, "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": 0.004185693757546613}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "rougeLsum_fmeasure": 0.19902709892722203, "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": 0.002712840280240763}, {"task_name": "GEM/web_nlg_en", "prompt_name": "very-explicit-description", "bleu": 0.945995909736822, "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": 0.050746289758330364}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.1948280360171628, "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": 0.0019190764140445665}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_recall": 0.35233986965204245, "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": 0.0027496772997846934}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_fmeasure": 0.23337611129812558, "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": 0.001789739809758051}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_precision": 0.045749458024920406, "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": 0.0008378390968457317}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_recall": 0.08650727583074418, "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": 0.0017368123677075}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_fmeasure": 0.05515519040890834, "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": 0.0009556470525659513}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_precision": 0.12931742952055322, "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": 0.001165133986135994}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_recall": 0.24476638208113655, "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": 0.0021953662674830537}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_fmeasure": 0.15705948233725145, "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": 0.0011392694655224288}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_precision": 0.18021848820258218, "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": 0.0017676765814184558}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_recall": 0.32710499965940787, "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": 0.002593215520354348}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_fmeasure": 0.21612005434473094, "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": 0.0016538900302117444}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "bleu": 2.46736793648531, "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": 0.06725208824507234}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.14631006684623113, "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": 0.00202940553368696}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_recall": 0.23364997943846966, "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": 0.0029065790428781588}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_fmeasure": 0.16297547934323248, "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": 0.001933969051854882}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_precision": 0.025827847916941985, "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": 0.0007754709320357255}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_recall": 0.04526934919886777, "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": 0.0014203298871475897}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_fmeasure": 0.029765592510798185, "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": 0.0008324139100615708}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_precision": 0.10792226209896028, "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": 0.0014088233241247392}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_recall": 0.17617202449317115, "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": 0.0021723160479157383}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_fmeasure": 0.12043877865207586, "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": 0.0012870984340208548}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_precision": 0.13660717079420895, "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": 0.0018788917566384465}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_recall": 0.21825446404308613, "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": 0.0026976088321342584}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_fmeasure": 0.1520750131577038, "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": 0.0017796530367973282}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "bleu": 1.5477983498042156, "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": 0.050133489193120745}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.13079039759634417, "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": 0.001955543608782996}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_recall": 0.2055460175970155, "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": 0.0026660575151607035}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_fmeasure": 0.1439324703457309, "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": 0.0017726885367313095}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_precision": 0.02019255851252806, "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": 0.000701616398727989}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_recall": 0.03433692695661584, "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": 0.0012373126733162317}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_fmeasure": 0.022911092202916015, "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": 0.0007479054746410414}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_precision": 0.09999640686456816, "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": 0.001424126236432219}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_recall": 0.1605856514940703, "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": 0.0020193184103868334}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_fmeasure": 0.11044014098329956, "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": 0.0012351812560723733}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_precision": 0.12199889988644426, "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": 0.0018145494431735837}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_recall": 0.1917031580824146, "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": 0.002465395188397278}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_fmeasure": 0.13411167349622463, "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": 0.0016260571621142044}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "bleu": 1.147232298385691, "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": 0.0691826847795543}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.11327911282182182, "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": 0.0021050526028020817}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_recall": 0.16828180201610024, "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": 0.002847973637534851}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_fmeasure": 0.11856241158259083, "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": 0.0018809770570257767}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_precision": 0.017366152232548535, "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": 0.000688515696922413}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_recall": 0.02843022440850937, "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": 0.0012109345375360039}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_fmeasure": 0.018749436523803075, "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": 0.0006925636650620048}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_precision": 0.08921385332529877, "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": 0.0016211137571385916}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_recall": 0.13492630719990034, "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": 0.0022450982819197966}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_fmeasure": 0.09341747472671912, "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": 0.001386769494087163}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_precision": 0.10532122185416891, "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": 0.001964559705223317}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_recall": 0.15676138791304953, "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": 0.0026400834873639987}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_fmeasure": 0.11018879107712176, "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": 0.00173454590238903}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "bleu": 1.1684671537023807, "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": 0.04924834856380638}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.03778312043913817, "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": 0.0015926234737421885}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_recall": 0.05436751166293101, "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": 0.0021832790923792976}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_fmeasure": 0.03775075623623371, "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": 0.0014413255950182725}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_precision": 0.006005536291342319, "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": 0.00046211258725366414}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_recall": 0.010297281978903448, "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": 0.0008604766653648099}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_fmeasure": 0.006221366299003606, "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": 0.0004335186102878494}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_precision": 0.030810947169269184, "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": 0.0012752703922055816}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_recall": 0.044772829387249816, "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": 0.0017657756805190348}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_fmeasure": 0.03071129391249539, "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": 0.0011276825153294747}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_precision": 0.035147111067221175, "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": 0.001484208404098812}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_recall": 0.05025842551722301, "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": 0.0020068457531649342}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_fmeasure": 0.03494607422776613, "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": 0.0013247268190345314}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "bleu": 0.17595126883839032, "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": 0.015567070968680874}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.0074793507056978, "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": 0.0010053147685702064}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_recall": 0.008777082891510336, "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": 0.0009666426404550582}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge1_fmeasure": 0.005906404781985648, "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": 0.000607677107759658}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_precision": 0.0014030770957601634, "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": 0.00029828016681575804}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_recall": 0.0020412684734586305, "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": 0.0004802932361861635}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rouge2_fmeasure": 0.0011421300202245823, "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": 0.0001895633931157293}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_precision": 0.006391064218189315, "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": 0.0009109362133582975}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_recall": 0.007472582117091991, "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": 0.0008445154328235708}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeL_fmeasure": 0.004863362352523752, "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": 0.000483094699749852}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_precision": 0.007072556922535605, "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": 0.0009667368382235277}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_recall": 0.008316788398208642, "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": 0.0009283822986494528}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "rougeLsum_fmeasure": 0.005541652569745538, "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": 0.0005675610525601405}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "article_summary_en", "bleu": 3.408899462135777e-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": 7.753498169402559e-08}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.09192315162664168, "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": 0.0016451913606095115}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_recall": 0.13873404877019072, "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": 0.0023406529305017346}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_fmeasure": 0.10157220185830643, "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": 0.001652719682743329}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_precision": 0.010919751857460677, "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": 0.0004590218726372067}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_recall": 0.01818440275603619, "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": 0.0008821073322667799}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_fmeasure": 0.012456692758547655, "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": 0.0005219947141645256}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_precision": 0.07787848338985379, "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": 0.0012726592773474163}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_recall": 0.12047198485093578, "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": 0.0019642618829164545}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_fmeasure": 0.08668933960294702, "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": 0.0012862012023578274}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_precision": 0.08531196788821724, "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": 0.0015072005308535098}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_recall": 0.12925517955975716, "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": 0.0021672933213231775}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_fmeasure": 0.09435362980226468, "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": 0.0015146600528512554}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "bleu": 0.6570455142605232, "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": 0.039377236612739966}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.11232161995413663, "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": 0.0014986011473891842}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_recall": 0.1090254224946869, "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": 0.0014272907431896768}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_fmeasure": 0.09779627553379834, "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": 0.0011383833609530695}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_precision": 0.0044729066220370204, "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": 0.00028499954551561567}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_recall": 0.004086203576860392, "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": 0.00027239799931189594}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_fmeasure": 0.0037787016720193267, "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": 0.00023237472662339907}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_precision": 0.09051739614158358, "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": 0.0011785667329494089}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_recall": 0.08813434495086396, "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": 0.001135626192780101}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_fmeasure": 0.07837775454126551, "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": 0.0008534564170118882}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_precision": 0.10847659377396374, "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": 0.0014321466350022917}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_recall": 0.10557207297887106, "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": 0.0013789439854323857}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_fmeasure": 0.0945568916084436, "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": 0.0010905899894760752}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "bleu": 0.23011553663489218, "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": 0.023163457968695483}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.114701552958001, "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": 0.0015443973941061887}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_recall": 0.11817859995979002, "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": 0.0016745011366248277}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_fmeasure": 0.10238757087450429, "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": 0.0012357283726096779}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_precision": 0.0060647194475853565, "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": 0.0003652209139100122}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_recall": 0.006920781836415898, "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": 0.0004604508533110552}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_fmeasure": 0.005616982781388312, "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": 0.00031568073407139486}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_precision": 0.09190681009159374, "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": 0.0012108271636288377}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_recall": 0.09506979579792599, "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": 0.0013149659100426684}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_fmeasure": 0.08150407552351425, "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": 0.0009086031747064198}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_precision": 0.11030994096605791, "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": 0.0014745909810591179}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_recall": 0.11380605707789511, "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": 0.0016064824284080005}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_fmeasure": 0.09851962625247197, "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": 0.0011780463149380163}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "bleu": 0.42720449416766143, "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": 0.04920014489048216}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.10206821399556017, "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": 0.001844530915726009}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_recall": 0.10677109354716818, "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": 0.0020498058139481567}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_fmeasure": 0.08950173963867221, "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": 0.001489985499534353}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_precision": 0.007296654670130939, "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": 0.0004567673436650055}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_recall": 0.009678010443095885, "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": 0.0006380621154672373}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_fmeasure": 0.0071627598071059735, "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": 0.00041755441992553}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_precision": 0.08163712210782142, "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": 0.0014510578299749544}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_recall": 0.08589657299944661, "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": 0.001617504296847833}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_fmeasure": 0.07103105337068176, "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": 0.0011045655139066616}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_precision": 0.09687603660663303, "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": 0.0017332241844581423}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_recall": 0.10098270644961956, "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": 0.0019132300942898686}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_fmeasure": 0.08475070434730132, "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": 0.001388696169181361}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "bleu": 0.5798662703555891, "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": 0.054465355095600916}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.03533145896571231, "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": 0.0015549597647742288}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_recall": 0.03875492098389553, "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": 0.001801752682403263}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_fmeasure": 0.02985352782800046, "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": 0.0012188795214852677}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_precision": 0.004012146000978968, "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": 0.00042222579770029815}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_recall": 0.006096521359894151, "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": 0.0007235591877570245}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_fmeasure": 0.0036360323548543325, "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": 0.0003318798006380848}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_precision": 0.029548821287779272, "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": 0.0013233063784575828}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_recall": 0.032023305030984744, "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": 0.0014725942164208706}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_fmeasure": 0.02447971050802285, "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": 0.0009617845148751048}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_precision": 0.033551820098614536, "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": 0.001487061201477005}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_recall": 0.036400040049703285, "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": 0.001681140548489408}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_fmeasure": 0.028155928024541907, "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": 0.0011428461263464302}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "bleu": 0.05849581177993631, "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": 0.009180398240758107}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.006370780717142262, "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": 0.0007837057340974754}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_recall": 0.006411871782469116, "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": 0.0008239031642928888}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge1_fmeasure": 0.004939617845914015, "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": 0.0005578424736919035}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_precision": 0.0009938789002252382, "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": 0.000237999522034536}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_recall": 0.0011447829515024383, "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": 0.00037870686521238793}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rouge2_fmeasure": 0.000662685399307501, "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": 0.00013198270880311396}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_precision": 0.005367348005361629, "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": 0.0006680343487566941}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_recall": 0.005288382317613349, "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": 0.0006938184123459739}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeL_fmeasure": 0.004048280211560025, "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": 0.0004499096165513398}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_precision": 0.006204481703520406, "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": 0.0007714054815105551}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_recall": 0.006146820943589796, "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": 0.0007932590903090583}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "rougeLsum_fmeasure": 0.004748006339916844, "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": 0.0005350205478695825}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "rephrase_en", "bleu": 4.72400314707023e-12, "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": 4.559094168568337e-11}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.07612749173217778, "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": 0.0015195087489753472}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_recall": 0.11262337276381704, "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": 0.0019042687612725346}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_fmeasure": 0.08305491429552378, "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": 0.0014262545897216355}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_precision": 0.005957943553532386, "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": 0.00035656165264230243}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_recall": 0.009062062070233518, "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": 0.0006086920290379038}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_fmeasure": 0.006444798648883974, "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": 0.00036740962517373447}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_precision": 0.0687786326083349, "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": 0.0013031149778182302}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_recall": 0.1035834052108859, "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": 0.0017150647479145709}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_fmeasure": 0.07560959548913089, "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": 0.001242462318430728}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_precision": 0.07117143547623304, "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": 0.0014119392640919361}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_recall": 0.10594508053345607, "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": 0.0017868259030921344}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_fmeasure": 0.07777041390254269, "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": 0.0013212339450664587}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "bleu": 0.32923260187928105, "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": 0.018591945193941646}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.11328604076040243, "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": 0.001506732252081756}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_recall": 0.10965033757032606, "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": 0.0014377266777683994}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_fmeasure": 0.0985048783907421, "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": 0.001138486531854058}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_precision": 0.004584687640655778, "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": 0.0002860681322232273}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_recall": 0.0043467286955627075, "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": 0.0002913190661459278}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_fmeasure": 0.003919360864074615, "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": 0.0002362379316907881}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_precision": 0.09115954245373961, "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": 0.0011853574111024992}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_recall": 0.08845616894390658, "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": 0.0011399661391188743}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_fmeasure": 0.07881651156803333, "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": 0.0008531520266409964}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_precision": 0.10940122241594495, "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": 0.0014399624875810046}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_recall": 0.10614639974822696, "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": 0.0013857248236570287}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_fmeasure": 0.09523468339897702, "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": 0.0010910535400450194}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "bleu": 0.24267691815638562, "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": 0.023329990719428425}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.11855473811641502, "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": 0.0017362173213766056}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_recall": 0.12436180887260508, "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": 0.0017727470217501237}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_fmeasure": 0.10565827408801438, "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": 0.0012895122486018143}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_precision": 0.008502239381305222, "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": 0.0005915872573552433}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_recall": 0.009242056572048922, "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": 0.000609865192415655}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_fmeasure": 0.0071178788883793365, "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": 0.0003656351117532055}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_precision": 0.09536132542545382, "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": 0.0013775517304080065}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_recall": 0.1009972841005859, "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": 0.001426517240170896}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_fmeasure": 0.08468987974885644, "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": 0.0009615948889506701}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_precision": 0.11385871311241222, "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": 0.0016704897346133682}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_recall": 0.11915527557338061, "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": 0.001683236148110152}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_fmeasure": 0.10126003544098408, "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": 0.0012189160081999214}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "bleu": 0.5278549864111844, "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": 0.05398385462423686}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.10386388460600006, "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": 0.0020564326439674687}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_recall": 0.10977300126578513, "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": 0.0021092462223978987}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_fmeasure": 0.0906629757475286, "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": 0.0015394873960304554}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_precision": 0.009698453240029803, "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": 0.00058283063405479}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_recall": 0.012564124080892733, "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": 0.0007772973629584162}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_fmeasure": 0.00904767741246584, "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": 0.00045357127991255266}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_precision": 0.08501105510332313, "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": 0.0016782898106444664}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_recall": 0.09117975113175547, "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": 0.0017506103268174781}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_fmeasure": 0.07410143647677019, "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": 0.0011995283359921434}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_precision": 0.09858732940096966, "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": 0.0019529795307488362}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_recall": 0.10394759619102645, "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": 0.0019908045367016767}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_fmeasure": 0.08582296561721141, "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": 0.0014436138607594486}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "bleu": 0.7246714720678235, "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": 0.042996486770572356}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.03681873828827321, "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": 0.0017367730475760668}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_recall": 0.034717284628656205, "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": 0.001598041936480893}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_fmeasure": 0.028755759541199397, "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": 0.0012275615171437174}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_precision": 0.0050668631877150705, "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": 0.0006212762145824261}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_recall": 0.005214938032076526, "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": 0.0005831344522889276}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_fmeasure": 0.0038530980639557002, "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": 0.00039777720107802043}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_precision": 0.031187038542479232, "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": 0.0014904198899705126}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_recall": 0.02901029742811881, "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": 0.0013109807032808127}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_fmeasure": 0.023977752584340473, "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": 0.0010083675242275193}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_precision": 0.03510037089937253, "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": 0.001669629173927134}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_recall": 0.03272422922075702, "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": 0.0014981296646924122}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_fmeasure": 0.02718496564188339, "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": 0.001156406107584219}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "bleu": 0.053208301348650455, "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": 0.007982915918716955}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.004584567706903281, "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": 0.0005862172465713146}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_recall": 0.004883341698702832, "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": 0.0006918213340664053}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge1_fmeasure": 0.003882818003436804, "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": 0.0004744262306385224}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_precision": 0.0002963908437119633, "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": 8.2186426279902e-05}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_recall": 0.0007327405127788758, "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": 0.00035674123771197504}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rouge2_fmeasure": 0.000329776680399744, "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": 9.27191454102303e-05}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_precision": 0.0038091284832852497, "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": 0.0004685137812954889}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_recall": 0.004225009801319019, "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": 0.0006169855878604657}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeL_fmeasure": 0.0032566156365132835, "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": 0.00037958025650827027}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_precision": 0.0044193000541017284, "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": 0.0005668947851862237}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_recall": 0.004734801369567873, "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": 0.0006763802665077723}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "rougeLsum_fmeasure": 0.003735454864613756, "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": 0.00045384136701505384}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "summarize_above_en", "bleu": 9.884916149064335e-14, "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": 1.3828817079734034e-12}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.10931885935430803, "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": 0.00191871599373175}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rouge1_recall": 0.16023646968449579, "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": 0.002510864634092151}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rouge1_fmeasure": 0.11890757521473502, "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": 0.0018199624825625143}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rouge2_precision": 0.014867695750213985, "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": 0.0005673573269887104}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rouge2_recall": 0.023968108745063707, "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": 0.0010773274186447375}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rouge2_fmeasure": 0.01658834501923122, "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": 0.0006230863250652812}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rougeL_precision": 0.09135469007050026, "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": 0.0015005226730239116}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rougeL_recall": 0.13744249919824175, "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": 0.002121616217495362}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rougeL_fmeasure": 0.10027588363073844, "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": 0.0014362856885268165}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rougeLsum_precision": 0.10128688810185907, "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": 0.0017621931672045334}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rougeLsum_recall": 0.14933392567748432, "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": 0.002341932353472351}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rougeLsum_fmeasure": 0.1102947959271491, "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": 0.0016670927564840557}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "bleu": 0.9307097145043712, "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": 0.07677320534555479}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}
220m7b57b5/eval/agg.lm1-220m-7b5_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": 0.12423492015375089, "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": 0.0018292924952296881}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rouge1_recall": 0.14407993186649484, "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": 0.002237465354397232}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rouge1_fmeasure": 0.1159016170094812, "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": 0.001521344866083945}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rouge2_precision": 0.012248945488953555, "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": 0.0006335078870070797}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rouge2_recall": 0.01661626736947049, "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": 0.0008273498401762434}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rouge2_fmeasure": 0.01194452012557495, "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": 0.0005382370999604981}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rougeL_precision": 0.09908351565617876, "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": 0.0014172741591640193}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rougeL_recall": 0.11617391929703827, "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": 0.0017732914443391238}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rougeL_fmeasure": 0.0921956787228759, "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": 0.0011273610173859465}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rougeLsum_precision": 0.11809874436847956, "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": 0.0017189280192798347}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rougeLsum_recall": 0.13652875271224274, "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": 0.0020724551557288213}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "rougeLsum_fmeasure": 0.10993399327419623, "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": 0.0014045764184309924}, {"task_name": "GEM/wiki_lingua_en", "prompt_name": "tldr_en", "bleu": 0.8885542546818795, "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": 0.05948451637908914}], "config": {"model": "hf-causal", "model_args": "pretrained=/pfs/lustrep4/scratch/project_462000119/muennighoff/nov-2022-bettercom/lm1-220m-7b5/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}}