bootstrapping for stddev: perplexity
{
  "results": {
    "arc_challenge": {
      "acc,none": 0.18686006825938567,
      "acc_stderr,none": 0.011391015649694396,
      "acc_norm,none": 0.24146757679180889,
      "acc_norm_stderr,none": 0.012506564839739432
    },
    "arc_easy": {
      "acc,none": 0.45707070707070707,
      "acc_stderr,none": 0.010221897564256042,
      "acc_norm,none": 0.4335016835016835,
      "acc_norm_stderr,none": 0.010168640625454107
    },
    "boolq": {
      "acc,none": 0.5938837920489297,
      "acc_stderr,none": 0.008589510943787407
    },
    "hellaswag": {
      "acc,none": 0.2861979685321649,
      "acc_stderr,none": 0.004510593395289897,
      "acc_norm,none": 0.3037243576976698,
      "acc_norm_stderr,none": 0.00458925224362782
    },
    "lambada_openai": {
      "perplexity,none": 60.10678259260095,
      "perplexity_stderr,none": 2.70306190425745,
      "acc,none": 0.2829419755482243,
      "acc_stderr,none": 0.006275349431343604
    },
    "openbookqa": {
      "acc,none": 0.164,
      "acc_stderr,none": 0.016575811142446713,
      "acc_norm,none": 0.256,
      "acc_norm_stderr,none": 0.019536923574747598
    },
    "piqa": {
      "acc,none": 0.6349292709466812,
      "acc_stderr,none": 0.011233021830554834,
      "acc_norm,none": 0.6305767138193689,
      "acc_norm_stderr,none": 0.011260988628572333
    },
    "sciq": {
      "acc,none": 0.77,
      "acc_stderr,none": 0.013314551335935952,
      "acc_norm,none": 0.747,
      "acc_norm_stderr,none": 0.01375427861358708
    },
    "wikitext": {
      "word_perplexity,none": 69.5803814818009,
      "byte_perplexity,none": 2.012533549921474,
      "bits_per_byte,none": 1.009012833861732
    },
    "winogrande": {
      "acc,none": 0.516179952644041,
      "acc_stderr,none": 0.0140451261309786
    }
  },
  "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 0x7f50b9b54040>",
      "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 0x7f50b9b26ef0>",
      "doc_to_target": "<function doc_to_target at 0x7f50b9b27370>",
      "doc_to_choice": "<function doc_to_choice at 0x7f50b9b275b0>",
      "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-pythia160m-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-pythia160m-hh-sft), limit: None, num_fewshot: 5, batch_size: 16
|     Task     |Version|Filter|    Metric     | Value |   |Stderr|
|--------------|-------|------|---------------|------:|---|-----:|
|arc_challenge |Yaml   |none  |acc            | 0.1869|±  |0.0114|
|              |       |none  |acc_norm       | 0.2415|±  |0.0125|
|arc_easy      |Yaml   |none  |acc            | 0.4571|±  |0.0102|
|              |       |none  |acc_norm       | 0.4335|±  |0.0102|
|boolq         |Yaml   |none  |acc            | 0.5939|±  |0.0086|
|hellaswag     |Yaml   |none  |acc            | 0.2862|±  |0.0045|
|              |       |none  |acc_norm       | 0.3037|±  |0.0046|
|lambada_openai|Yaml   |none  |perplexity     |60.1068|±  |2.7031|
|              |       |none  |acc            | 0.2829|±  |0.0063|
|openbookqa    |Yaml   |none  |acc            | 0.1640|±  |0.0166|
|              |       |none  |acc_norm       | 0.2560|±  |0.0195|
|piqa          |Yaml   |none  |acc            | 0.6349|±  |0.0112|
|              |       |none  |acc_norm       | 0.6306|±  |0.0113|
|sciq          |Yaml   |none  |acc            | 0.7700|±  |0.0133|
|              |       |none  |acc_norm       | 0.7470|±  |0.0138|
|wikitext      |Yaml   |none  |word_perplexity|69.5804|   |      |
|              |       |none  |byte_perplexity| 2.0125|   |      |
|              |       |none  |bits_per_byte  | 1.0090|   |      |
|winogrande    |Yaml   |none  |acc            | 0.5162|±  |0.0140|