eleuther-pythia70m-hh-sft / sft-70m-eval-files /sft-pythia-70m-5shot-shelloutput.txt
lomahony's picture
Upload 8 files
aaae80f
bootstrapping for stddev: perplexity
{
"results": {
"arc_challenge": {
"acc,none": 0.1757679180887372,
"acc_stderr,none": 0.011122850863120487,
"acc_norm,none": 0.21245733788395904,
"acc_norm_stderr,none": 0.011953482906582947
},
"arc_easy": {
"acc,none": 0.38173400673400676,
"acc_stderr,none": 0.00996864885183967,
"acc_norm,none": 0.3661616161616162,
"acc_norm_stderr,none": 0.00988539139094772
},
"boolq": {
"acc,none": 0.5837920489296636,
"acc_stderr,none": 0.008621380519419278
},
"hellaswag": {
"acc,none": 0.2605058753236407,
"acc_stderr,none": 0.004380136468543921,
"acc_norm,none": 0.2665803624775941,
"acc_norm_stderr,none": 0.004412674170976497
},
"lambada_openai": {
"perplexity,none": 495.31700918494755,
"perplexity_stderr,none": 27.7861411859479,
"acc,none": 0.15854841839705025,
"acc_stderr,none": 0.005088708755423655
},
"openbookqa": {
"acc,none": 0.132,
"acc_stderr,none": 0.015152927850580162,
"acc_norm,none": 0.244,
"acc_norm_stderr,none": 0.019226734893614584
},
"piqa": {
"acc,none": 0.5870511425462459,
"acc_stderr,none": 0.011487658725079094,
"acc_norm,none": 0.5930359085963003,
"acc_norm_stderr,none": 0.011462093919190166
},
"sciq": {
"acc,none": 0.621,
"acc_stderr,none": 0.015349091002225347,
"acc_norm,none": 0.576,
"acc_norm_stderr,none": 0.015635487471405186
},
"wikitext": {
"word_perplexity,none": 180.27069075802413,
"byte_perplexity,none": 2.3545097616343265,
"bits_per_byte,none": 1.2354267039635143
},
"winogrande": {
"acc,none": 0.5209155485398579,
"acc_stderr,none": 0.014040185494212949
}
},
"configs": {
"arc_challenge": {
"task": "arc_challenge",
"group": [
"ai2_arc",
"multiple_choice"
],
"dataset_path": "ai2_arc",
"dataset_name": "ARC-Challenge",
"training_split": "train",
"validation_split": "validation",
"test_split": "test",
"doc_to_text": "Question: {{question}}\nAnswer:",
"doc_to_target": "{{choices.label.index(answerKey)}}",
"doc_to_choice": "{{choices.text}}",
"description": "",
"target_delimiter": " ",
"fewshot_delimiter": "\n\n",
"num_fewshot": 5,
"metric_list": [
{
"metric": "acc",
"aggregation": "mean",
"higher_is_better": true
},
{
"metric": "acc_norm",
"aggregation": "mean",
"higher_is_better": true
}
],
"output_type": "multiple_choice",
"repeats": 1,
"should_decontaminate": true,
"doc_to_decontamination_query": "Question: {{question}}\nAnswer:"
},
"arc_easy": {
"task": "arc_easy",
"group": [
"ai2_arc",
"multiple_choice"
],
"dataset_path": "ai2_arc",
"dataset_name": "ARC-Easy",
"training_split": "train",
"validation_split": "validation",
"test_split": "test",
"doc_to_text": "Question: {{question}}\nAnswer:",
"doc_to_target": "{{choices.label.index(answerKey)}}",
"doc_to_choice": "{{choices.text}}",
"description": "",
"target_delimiter": " ",
"fewshot_delimiter": "\n\n",
"num_fewshot": 5,
"metric_list": [
{
"metric": "acc",
"aggregation": "mean",
"higher_is_better": true
},
{
"metric": "acc_norm",
"aggregation": "mean",
"higher_is_better": true
}
],
"output_type": "multiple_choice",
"repeats": 1,
"should_decontaminate": true,
"doc_to_decontamination_query": "Question: {{question}}\nAnswer:"
},
"boolq": {
"task": "boolq",
"group": [
"super-glue-lm-eval-v1"
],
"dataset_path": "super_glue",
"dataset_name": "boolq",
"training_split": "train",
"validation_split": "validation",
"doc_to_text": "{{passage}}\nQuestion: {{question}}?\nAnswer:",
"doc_to_target": "label",
"doc_to_choice": [
"no",
"yes"
],
"description": "",
"target_delimiter": " ",
"fewshot_delimiter": "\n\n",
"num_fewshot": 5,
"metric_list": [
{
"metric": "acc"
}
],
"output_type": "multiple_choice",
"repeats": 1,
"should_decontaminate": true,
"doc_to_decontamination_query": "passage"
},
"hellaswag": {
"task": "hellaswag",
"group": [
"multiple_choice"
],
"dataset_path": "hellaswag",
"training_split": "train",
"validation_split": "validation",
"doc_to_text": "{% set text = activity_label ~ ': ' ~ ctx_a ~ ' ' ~ ctx_b.capitalize() %}{{text|trim|replace(' [title]', '. ')|regex_replace('\\[.*?\\]', '')|replace(' ', ' ')}}",
"doc_to_target": "{{label}}",
"doc_to_choice": "{{endings|map('trim')|map('replace', ' [title]', '. ')|map('regex_replace', '\\[.*?\\]', '')|map('replace', ' ', ' ')|list}}",
"description": "",
"target_delimiter": " ",
"fewshot_delimiter": "\n\n",
"num_fewshot": 5,
"metric_list": [
{
"metric": "acc",
"aggregation": "mean",
"higher_is_better": true
},
{
"metric": "acc_norm",
"aggregation": "mean",
"higher_is_better": true
}
],
"output_type": "multiple_choice",
"repeats": 1,
"should_decontaminate": false
},
"lambada_openai": {
"task": "lambada_openai",
"group": [
"lambada",
"loglikelihood",
"perplexity"
],
"dataset_path": "EleutherAI/lambada_openai",
"dataset_name": "default",
"test_split": "test",
"template_aliases": "",
"doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
"doc_to_target": "{{' '+text.split(' ')[-1]}}",
"description": "",
"target_delimiter": " ",
"fewshot_delimiter": "\n\n",
"num_fewshot": 5,
"metric_list": [
{
"metric": "perplexity",
"aggregation": "perplexity",
"higher_is_better": false
},
{
"metric": "acc",
"aggregation": "mean",
"higher_is_better": true
}
],
"output_type": "loglikelihood",
"repeats": 1,
"should_decontaminate": true,
"doc_to_decontamination_query": "{{text}}"
},
"openbookqa": {
"task": "openbookqa",
"group": [
"multiple_choice"
],
"dataset_path": "openbookqa",
"dataset_name": "main",
"training_split": "train",
"validation_split": "validation",
"test_split": "test",
"doc_to_text": "question_stem",
"doc_to_target": "{{choices.label.index(answerKey.lstrip())}}",
"doc_to_choice": "{{choices.text}}",
"description": "",
"target_delimiter": " ",
"fewshot_delimiter": "\n\n",
"num_fewshot": 5,
"metric_list": [
{
"metric": "acc",
"aggregation": "mean",
"higher_is_better": true
},
{
"metric": "acc_norm",
"aggregation": "mean",
"higher_is_better": true
}
],
"output_type": "multiple_choice",
"repeats": 1,
"should_decontaminate": true,
"doc_to_decontamination_query": "question_stem"
},
"piqa": {
"task": "piqa",
"group": [
"multiple_choice"
],
"dataset_path": "piqa",
"training_split": "train",
"validation_split": "validation",
"doc_to_text": "Question: {{goal}}\nAnswer:",
"doc_to_target": "label",
"doc_to_choice": "{{[sol1, sol2]}}",
"description": "",
"target_delimiter": " ",
"fewshot_delimiter": "\n\n",
"num_fewshot": 5,
"metric_list": [
{
"metric": "acc",
"aggregation": "mean",
"higher_is_better": true
},
{
"metric": "acc_norm",
"aggregation": "mean",
"higher_is_better": true
}
],
"output_type": "multiple_choice",
"repeats": 1,
"should_decontaminate": true,
"doc_to_decontamination_query": "goal"
},
"sciq": {
"task": "sciq",
"group": [
"multiple_choice"
],
"dataset_path": "sciq",
"training_split": "train",
"validation_split": "validation",
"test_split": "test",
"doc_to_text": "{{support.lstrip()}}\nQuestion: {{question}}\nAnswer:",
"doc_to_target": 3,
"doc_to_choice": "{{[distractor1, distractor2, distractor3, correct_answer]}}",
"description": "",
"target_delimiter": " ",
"fewshot_delimiter": "\n\n",
"num_fewshot": 5,
"metric_list": [
{
"metric": "acc",
"aggregation": "mean",
"higher_is_better": true
},
{
"metric": "acc_norm",
"aggregation": "mean",
"higher_is_better": true
}
],
"output_type": "multiple_choice",
"repeats": 1,
"should_decontaminate": true,
"doc_to_decontamination_query": "{{support}} {{question}}"
},
"wikitext": {
"task": "wikitext",
"group": [
"perplexity",
"loglikelihood_rolling"
],
"dataset_path": "EleutherAI/wikitext_document_level",
"dataset_name": "wikitext-2-raw-v1",
"training_split": "train",
"validation_split": "validation",
"test_split": "test",
"template_aliases": "",
"doc_to_text": "",
"doc_to_target": "<function wikitext_detokenizer at 0x7f54329bc040>",
"description": "",
"target_delimiter": " ",
"fewshot_delimiter": "\n\n",
"num_fewshot": 5,
"metric_list": [
{
"metric": "word_perplexity"
},
{
"metric": "byte_perplexity"
},
{
"metric": "bits_per_byte"
}
],
"output_type": "loglikelihood_rolling",
"repeats": 1,
"should_decontaminate": true,
"doc_to_decontamination_query": "{{page}}"
},
"winogrande": {
"task": "winogrande",
"dataset_path": "winogrande",
"dataset_name": "winogrande_xl",
"training_split": "train",
"validation_split": "validation",
"doc_to_text": "<function doc_to_text at 0x7f543298eef0>",
"doc_to_target": "<function doc_to_target at 0x7f543298f370>",
"doc_to_choice": "<function doc_to_choice at 0x7f543298f5b0>",
"description": "",
"target_delimiter": " ",
"fewshot_delimiter": "\n\n",
"num_fewshot": 5,
"metric_list": [
{
"metric": "acc",
"aggregation": "mean",
"higher_is_better": true
}
],
"output_type": "multiple_choice",
"repeats": 1,
"should_decontaminate": false
}
},
"versions": {
"arc_challenge": "Yaml",
"arc_easy": "Yaml",
"boolq": "Yaml",
"hellaswag": "Yaml",
"lambada_openai": "Yaml",
"openbookqa": "Yaml",
"piqa": "Yaml",
"sciq": "Yaml",
"wikitext": "Yaml",
"winogrande": "Yaml"
},
"config": {
"model": "hf",
"model_args": "pretrained=lomahony/eleuther-pythia70m-hh-sft",
"num_fewshot": 5,
"batch_size": 16,
"batch_sizes": [],
"device": "cuda:0",
"use_cache": null,
"limit": null,
"bootstrap_iters": 100000
},
"git_hash": "4e44f0a"
}
hf (pretrained=lomahony/eleuther-pythia70m-hh-sft), limit: None, num_fewshot: 5, batch_size: 16
| Task |Version|Filter| Metric | Value | |Stderr |
|--------------|-------|------|---------------|-------:|---|------:|
|arc_challenge |Yaml |none |acc | 0.1758|± | 0.0111|
| | |none |acc_norm | 0.2125|± | 0.0120|
|arc_easy |Yaml |none |acc | 0.3817|± | 0.0100|
| | |none |acc_norm | 0.3662|± | 0.0099|
|boolq |Yaml |none |acc | 0.5838|± | 0.0086|
|hellaswag |Yaml |none |acc | 0.2605|± | 0.0044|
| | |none |acc_norm | 0.2666|± | 0.0044|
|lambada_openai|Yaml |none |perplexity |495.3170|± |27.7861|
| | |none |acc | 0.1585|± | 0.0051|
|openbookqa |Yaml |none |acc | 0.1320|± | 0.0152|
| | |none |acc_norm | 0.2440|± | 0.0192|
|piqa |Yaml |none |acc | 0.5871|± | 0.0115|
| | |none |acc_norm | 0.5930|± | 0.0115|
|sciq |Yaml |none |acc | 0.6210|± | 0.0153|
| | |none |acc_norm | 0.5760|± | 0.0156|
|wikitext |Yaml |none |word_perplexity|180.2707| | |
| | |none |byte_perplexity| 2.3545| | |
| | |none |bits_per_byte | 1.2354| | |
|winogrande |Yaml |none |acc | 0.5209|± | 0.0140|