bootstrapping for stddev: perplexity
{
  "results": {
    "arc_challenge": {
      "acc,none": 0.25426621160409557,
      "acc_stderr,none": 0.012724999945157746,
      "acc_norm,none": 0.2883959044368601,
      "acc_norm_stderr,none": 0.013238394422428173
    },
    "arc_easy": {
      "acc,none": 0.5732323232323232,
      "acc_stderr,none": 0.010149141043955643,
      "acc_norm,none": 0.5542929292929293,
      "acc_norm_stderr,none": 0.010199118183322989
    },
    "boolq": {
      "acc,none": 0.5715596330275229,
      "acc_stderr,none": 0.008655028561519784
    },
    "hellaswag": {
      "acc,none": 0.3482374029077873,
      "acc_stderr,none": 0.004754380554929206,
      "acc_norm,none": 0.41236805417247563,
      "acc_norm_stderr,none": 0.00491254704013287
    },
    "lambada_openai": {
      "perplexity,none": 14.093662143573109,
      "perplexity_stderr,none": 0.48659081977443114,
      "acc,none": 0.42907044440131964,
      "acc_stderr,none": 0.006895529739245138
    },
    "openbookqa": {
      "acc,none": 0.19,
      "acc_stderr,none": 0.017561800410758988,
      "acc_norm,none": 0.284,
      "acc_norm_stderr,none": 0.020186703693570857
    },
    "piqa": {
      "acc,none": 0.6833514689880305,
      "acc_stderr,none": 0.010853160531978481,
      "acc_norm,none": 0.6828073993471164,
      "acc_norm_stderr,none": 0.01085815545438087
    },
    "sciq": {
      "acc,none": 0.864,
      "acc_stderr,none": 0.010845350230472988,
      "acc_norm,none": 0.85,
      "acc_norm_stderr,none": 0.011297239823409314
    },
    "wikitext": {
      "word_perplexity,none": 38.90357956698245,
      "byte_perplexity,none": 1.8285957273924667,
      "bits_per_byte,none": 0.8707361540620543
    },
    "winogrande": {
      "acc,none": 0.5272296764009471,
      "acc_stderr,none": 0.014031631629827694
    }
  },
  "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 0x7fd4f5060040>",
      "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 0x7fd4f5032ef0>",
      "doc_to_target": "<function doc_to_target at 0x7fd4f5033370>",
      "doc_to_choice": "<function doc_to_choice at 0x7fd4f50335b0>",
      "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-pythia410m-hh-dpo",
    "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-pythia410m-hh-dpo), limit: None, num_fewshot: 5, batch_size: 16
|     Task     |Version|Filter|    Metric     | Value |   |Stderr|
|--------------|-------|------|---------------|------:|---|-----:|
|arc_challenge |Yaml   |none  |acc            | 0.2543|±  |0.0127|
|              |       |none  |acc_norm       | 0.2884|±  |0.0132|
|arc_easy      |Yaml   |none  |acc            | 0.5732|±  |0.0101|
|              |       |none  |acc_norm       | 0.5543|±  |0.0102|
|boolq         |Yaml   |none  |acc            | 0.5716|±  |0.0087|
|hellaswag     |Yaml   |none  |acc            | 0.3482|±  |0.0048|
|              |       |none  |acc_norm       | 0.4124|±  |0.0049|
|lambada_openai|Yaml   |none  |perplexity     |14.0937|±  |0.4866|
|              |       |none  |acc            | 0.4291|±  |0.0069|
|openbookqa    |Yaml   |none  |acc            | 0.1900|±  |0.0176|
|              |       |none  |acc_norm       | 0.2840|±  |0.0202|
|piqa          |Yaml   |none  |acc            | 0.6834|±  |0.0109|
|              |       |none  |acc_norm       | 0.6828|±  |0.0109|
|sciq          |Yaml   |none  |acc            | 0.8640|±  |0.0108|
|              |       |none  |acc_norm       | 0.8500|±  |0.0113|
|wikitext      |Yaml   |none  |word_perplexity|38.9036|   |      |
|              |       |none  |byte_perplexity| 1.8286|   |      |
|              |       |none  |bits_per_byte  | 0.8707|   |      |
|winogrande    |Yaml   |none  |acc            | 0.5272|±  |0.0140|

