{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/truthfulqa-measuring-how-models-mimic-human","title":"TruthfulQA: Measuring How Models Mimic Human Falsehoods","arxiv_id":"2109.07958","date":"2021-09-08","proceeding":"ACL 2022 5","authors":["Stephanie Lin","Jacob Hilton","Owain Evans"],"abstract":"We propose a benchmark to measure whether a language model is truthful in generating answers to questions. The benchmark comprises 817 questions that span 38 categories, including health, law, finance and politics. We crafted questions that some humans would answer falsely due to a false belief or misconception. To perform well, models must avoid generating false answers learned from imitating human texts. We tested GPT-3, GPT-Neo/J, GPT-2 and a T5-based model. The best model was truthful on 58% of questions, while human performance was 94%. Models generated many false answers that mimic popular misconceptions and have the potential to deceive humans. The largest models were generally the least truthful. This contrasts with other NLP tasks, where performance improves with model size. However, this result is expected if false answers are learned from the training distribution. We suggest that scaling up models alone is less promising for improving truthfulness than fine-tuning using training objectives other than imitation of text from the web.","url_abs":"https://arxiv.org/abs/2109.07958v2","url_pdf":"https://arxiv.org/pdf/2109.07958v2.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"truthfulqa-measuring-how-models-mimic-human","repo_url":"https://github.com/sylinrl/truthfulqa","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"Apache-2.0"}},{"paper_slug":"truthfulqa-measuring-how-models-mimic-human","repo_url":"https://github.com/lurosenb/sass","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}},{"paper_slug":"truthfulqa-measuring-how-models-mimic-human","repo_url":"https://github.com/yizhongw/truthfulqa_reeval","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"Apache-2.0"}}],"tasks":[{"task_slug":"language-modeling","task_name":"Language Modeling"},{"task_slug":"language-modelling","task_name":"Language Modelling"},{"task_slug":"misconceptions","task_name":"Misconceptions"},{"task_slug":"question-answering","task_name":"Question Answering"},{"task_slug":"question-generation","task_name":"Question Generation"},{"task_slug":"truthfulqa","task_name":"TruthfulQA"}],"methods":[{"method_slug":"adam","method_name":"Adam"},{"method_slug":"attention","method_name":"Attention"},{"method_slug":"attention-dropout","method_name":"Attention Dropout"},{"method_slug":"bpe","method_name":"BPE"},{"method_slug":"cosine-annealing","method_name":"Cosine Annealing"},{"method_slug":"dense-connections","method_name":"Dense Connections"},{"method_slug":"discriminative-fine-tuning","method_name":"Discriminative Fine-Tuning"},{"method_slug":"dropout","method_name":"Dropout"},{"method_slug":"gpt-2","method_name":"GPT-2"},{"method_slug":"gpt-3","method_name":"GPT-3"},{"method_slug":"layer-normalization","method_name":"Layer Normalization"},{"method_slug":"linear-layer","method_name":"Linear Layer"},{"method_slug":"linear-warmup-with-cosine-annealing","method_name":"Linear Warmup With Cosine Annealing"},{"method_slug":"multi-head-attention","method_name":"Multi-Head Attention"},{"method_slug":"residual-connection","method_name":"Residual Connection"},{"method_slug":"softmax","method_name":"Softmax"},{"method_slug":"weight-decay","method_name":"Weight Decay"}],"datasets_introduced":[{"slug":"truthfulqa","name":"TruthfulQA","full_name":""}],"methods_introduced":[],"results":[{"leaderboard":"/sota/question-answering-on-truthfulqa","task":"Question Answering","dataset":"TruthfulQA","model":"GPT-2 1.5B","rank_in_archive_order":15,"of":33,"metrics":{"% info":"89.84","% true":"29.50","% true (GPT-judge)":"29.87","BLEU":"-4.91","BLEURT":"-0.25","MC1":"0.22","MC2":"0.39","ROUGE":"-9.41"},"uses_additional_data":false},{"leaderboard":"/sota/question-answering-on-truthfulqa","task":"Question Answering","dataset":"TruthfulQA","model":"GPT-3 175B","rank_in_archive_order":17,"of":33,"metrics":{"% info":"97.55","% true":"20.44","% true (GPT-judge)":"20.56","BLEU":"-17.38","BLEURT":"-0.56","MC1":"0.21","MC2":"0.33","ROUGE":"-17.75"},"uses_additional_data":false},{"leaderboard":"/sota/question-answering-on-truthfulqa","task":"Question Answering","dataset":"TruthfulQA","model":"GPT-J 6B","rank_in_archive_order":19,"of":33,"metrics":{"% info":"89.96","% true":"26.68","% true (GPT-judge)":"27.17","BLEU":"-7.58","BLEURT":"-0.31","MC1":"0.20","MC2":"0.36","ROUGE":"-11.35"},"uses_additional_data":false},{"leaderboard":"/sota/question-answering-on-truthfulqa","task":"Question Answering","dataset":"TruthfulQA","model":"UnifiedQA 3B","rank_in_archive_order":20,"of":33,"metrics":{"% info":"64.50","% true":"53.86","% true (GPT-judge)":"53.24","BLEU":"-0.16","BLEURT":"0.08","MC1":"0.19","MC2":"0.35","ROUGE":"1.76"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2109.07958","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2109.07958"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/lurosenb/sass","reach":{"status":"ok"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/yizhongw/truthfulqa_reeval","reach":{"status":"ok","spdx":"Apache-2.0"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/sylinrl/truthfulqa","reach":{"status":"ok","spdx":"Apache-2.0"}}],"summary":{"ran_draft_wrong":2},"by_repo_kind":{"official":{"samples":2,"ran":2,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":0,"samples":[{"code_sha256_prefix":"679c0af24337099d","entry":"data_to_dict","repo":"sylinrl/truthfulqa","repo_kind":"official","path":"truthfulqa/evaluate.py","file_url":"https://github.com/sylinrl/truthfulqa/blob/HEAD/truthfulqa/evaluate.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"MISDECLARED","metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"679c0af24337099d"}},{"code_sha256_prefix":"18c6154c3dcf6170","entry":"format_frame","repo":"sylinrl/truthfulqa","repo_kind":"official","path":"truthfulqa/evaluate.py","file_url":"https://github.com/sylinrl/truthfulqa/blob/HEAD/truthfulqa/evaluate.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"18c6154c3dcf6170"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}