{"about":{"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.","site":"https://codewithpapers.app","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","syntology":{"site":"https://syntology.ai","developers":"https://syntology.ai/developers","mcp":{"server":"https://syntology.ai/mcp","transport":"streamable-http","server_card":"https://syntology.ai/.well-known/mcp/server-card.json","auth":{"type":"trial token, no account","trial_token":"https://syntology.ai/api/oauth/trial/token","method":"POST","docs":"https://syntology.ai/developers"}},"have":"https://syntology.ai/api/graph/have?x=<method, arXiv id or title> (free, answers coverage only)","paper_base":"https://syntology.ai/paper/","atlas_base":"https://app.syntology.ai/?focus="},"machine_readable":[{"url":"https://codewithpapers.app/llms.txt","what":"the machine catalog: every machine-readable file, counted"},{"url":"https://codewithpapers.app/index/manifest.json","what":"paper-to-code index by arXiv id, with Syntology's counts"},{"url":"https://codewithpapers.app/search/manifest.json","what":"site search index (titles, authors) and its files"},{"url":"https://codewithpapers.app/download","what":"bulk files: Syntology's layer, described there"},{"url":"https://codewithpapers.app/build_manifest.json","what":"the build record: inputs, counts, exclusions, probes"}]},"url":"/task/truthfulqa","name":"TruthfulQA","slug":"truthfulqa","description_markdown":null,"categories":[{"name":"Knowledge Base","url":"/area/knowledge-base"}],"source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","slug_source":"archive_url"},"counts":{"papers_tagged":80,"papers_with_code":37,"benchmarks":0,"benchmark_tables_in_archive":0,"benchmark_tables_shown":0,"benchmark_tables_withheld_as_spam":0,"benchmark_definition":"a leaderboard table with at least one row; benchmark_tables_shown also counts the zero-row tables; benchmark_tables_in_archive adds the tables withheld as spam","datasets":3,"subtasks":0,"parent_tasks":0},"benchmarks":[],"datasets":[{"url":"/dataset/knowledge","name":"Knowledge","full_name":"","num_papers_in_archive":3},{"url":"/dataset/genaipabench-dataset","name":"GenAIPABench-Dataset","full_name":"","num_papers_in_archive":1},{"url":"/dataset/news","name":"News","full_name":"","num_papers_in_archive":1}],"subtasks":[],"parent_tasks":[],"papers":{"order":"repositories listed in the archive (most first), then date (newest first), then slug; the archive holds no stars","population":"every paper the archive tags with this task (distinct papers; spam titles not listed). 'of' counts the papers listed; until 2026-09-25 it counted only the papers with a code link, and that number is of_with_code","page":1,"pages_in_order":1,"rows_per_page":100,"shown":80,"of":80,"of_with_code":37,"tagged_in_all":80,"counts":{"archive_papers_tagged":80,"with_a_code_link":37,"where_syntology_ran_a_sample":20,"not_listed_spam_title":0,"listed":80,"listed_where_code_ran":20,"where_syntology_ran_a_sample_split":{"with_a_run_with_no_instrument_failure":14,"every_run_a_failure_of_syntologys_instrument":6,"listed_with_a_run_with_no_instrument_failure":14,"listed_every_run_a_failure_of_syntologys_instrument":6,"filter":{"states":["a run with no instrument failure","any run, instrument failures included"],"default":"a run with no instrument failure","note":"on the 'only where code ran' pages the default hides, in the browser, the rows where every run was a failure of Syntology's instrument; the second state shows them again. Rows are hidden, never re-ordered; these twins list every row"}},"definition":"distinct papers the archive tags; 'where Syntology ran a sample' counts papers with at least one harvested sample that ran, which is not a correctness claim"},"items":[{"url":"/paper/dola-decoding-by-contrasting-layers-improves","slug":"dola-decoding-by-contrasting-layers-improves","title":"DoLa: Decoding by Contrasting Layers Improves Factuality in Large Language Models","date":"2023-09-07","arxiv_id":"2309.03883","repositories_listed":4,"syntology":{"n":7,"n_ran":5,"n_constructed":2,"n_ran_checked":3,"n_instrument":2,"n_unverified":2,"n_honours":0,"n_violates":0,"n_no_contract":3,"n_pointer_only":7,"phrase":"5 ran (of which 2 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 2 where Syntology's instrument failed) · 2 unverified","sample_list":"/paper/dola-decoding-by-contrasting-layers-improves#ran","syntology_url":"https://syntology.ai/paper/2309.03883","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2309.03883"}},"official":{"repos":["voidism/dola"],"state":"official (archive's flag): 5 ran","n_ran":5,"n_constructed":2,"n_ran_no_instrument_failure":3,"n_unverified":2,"ran_from_kinds":["official"]}}},{"url":"/paper/rlhf-workflow-from-reward-modeling-to-online","slug":"rlhf-workflow-from-reward-modeling-to-online","title":"RLHF Workflow: From Reward Modeling to Online RLHF","date":"2024-05-13","arxiv_id":"2405.07863","repositories_listed":3,"syntology":{"n":1,"n_ran":0,"n_constructed":0,"n_ran_checked":0,"n_instrument":0,"n_unverified":1,"n_honours":0,"n_violates":0,"n_no_contract":0,"n_pointer_only":1,"phrase":"0 ran · 1 unverified","sample_list":"/paper/rlhf-workflow-from-reward-modeling-to-online#ran","syntology_url":"https://syntology.ai/paper/2405.07863","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2405.07863"}},"official":{"repos":["RLHFlow/RLHF-Reward-Modeling","rlhflow/online-rlhf"],"state":"official: harvested, nothing ran","n_ran":0,"n_constructed":0,"n_ran_no_instrument_failure":0,"n_unverified":1,"ran_from_kinds":[]}}},{"url":"/paper/truthfulqa-measuring-how-models-mimic-human","slug":"truthfulqa-measuring-how-models-mimic-human","title":"TruthfulQA: Measuring How Models Mimic Human Falsehoods","date":"2021-09-08","arxiv_id":"2109.07958","repositories_listed":3,"syntology":{"n":2,"n_ran":2,"n_constructed":0,"n_ran_checked":0,"n_instrument":2,"n_unverified":0,"n_honours":0,"n_violates":0,"n_no_contract":0,"n_pointer_only":0,"phrase":"2 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 2 where Syntology's instrument failed) · 0 unverified","sample_list":"/paper/truthfulqa-measuring-how-models-mimic-human#ran","syntology_url":"https://syntology.ai/paper/2109.07958","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2109.07958"}},"official":{"repos":["sylinrl/truthfulqa"],"state":"official (archive's flag): 2 ran","n_ran":2,"n_constructed":0,"n_ran_no_instrument_failure":0,"n_unverified":0,"ran_from_kinds":["official"]}}},{"url":"/paper/lacie-listener-aware-finetuning-for","slug":"lacie-listener-aware-finetuning-for","title":"LACIE: Listener-Aware Finetuning for Confidence Calibration in Large Language Models","date":"2024-05-31","arxiv_id":"2405.21028","repositories_listed":2,"syntology":null},{"url":"/paper/tuning-language-models-by-proxy","slug":"tuning-language-models-by-proxy","title":"Tuning Language Models by Proxy","date":"2024-01-16","arxiv_id":"2401.08565","repositories_listed":2,"syntology":{"n":10,"n_ran":9,"n_constructed":0,"n_ran_checked":4,"n_instrument":5,"n_unverified":1,"n_honours":1,"n_violates":0,"n_no_contract":3,"n_pointer_only":10,"phrase":"9 ran (of which 0 constructed an object rather than computing a result; 4 with no instrument failure: 1 honoured, 0 violated, 3 with no contract checked; 5 where Syntology's instrument failed) · 1 unverified","sample_list":"/paper/tuning-language-models-by-proxy#ran","syntology_url":"https://syntology.ai/paper/2401.08565","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2401.08565"}},"official":{"repos":["alisawuffles/proxy-tuning"],"state":"official (archive's flag): 9 ran","n_ran":9,"n_constructed":0,"n_ran_no_instrument_failure":4,"n_unverified":1,"ran_from_kinds":["official"]}}},{"url":"/paper/alleviating-hallucinations-of-large-language","slug":"alleviating-hallucinations-of-large-language","title":"Alleviating Hallucinations of Large Language Models through Induced Hallucinations","date":"2023-12-25","arxiv_id":"2312.15710","repositories_listed":2,"syntology":{"n":13,"n_ran":7,"n_constructed":0,"n_ran_checked":2,"n_instrument":5,"n_unverified":6,"n_honours":0,"n_violates":0,"n_no_contract":2,"n_pointer_only":6,"phrase":"7 ran (of which 0 constructed an object rather than computing a result; 2 with no instrument failure: 0 honoured, 0 violated, 2 with no contract checked; 5 where Syntology's instrument failed) · 6 unverified","sample_list":"/paper/alleviating-hallucinations-of-large-language#ran","syntology_url":"https://syntology.ai/paper/2312.15710","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2312.15710"}},"official":{"repos":["hiyouga/llama-factory","hillzhang1999/icd"],"state":"official (archive's flag): 7 ran","n_ran":7,"n_constructed":0,"n_ran_no_instrument_failure":2,"n_unverified":6,"ran_from_kinds":["official"]}}},{"url":"/paper/red-teaming-large-language-models-using-chain","slug":"red-teaming-large-language-models-using-chain","title":"Red-Teaming Large Language Models using Chain of Utterances for Safety-Alignment","date":"2023-08-18","arxiv_id":"2308.09662","repositories_listed":2,"syntology":null},{"url":"/paper/unsupervised-elicitation-of-language-models","slug":"unsupervised-elicitation-of-language-models","title":"Unsupervised Elicitation of Language Models","date":"2025-06-11","arxiv_id":"2506.10139","repositories_listed":1,"syntology":null},{"url":"/paper/truth-neurons","slug":"truth-neurons","title":"Truth Neurons","date":"2025-05-18","arxiv_id":"2505.12182","repositories_listed":1,"syntology":null},{"url":"/paper/delta-a-decoding-strategy-based-on-logit","slug":"delta-a-decoding-strategy-based-on-logit","title":"DeLTa: A Decoding Strategy based on Logit Trajectory Prediction Improves Factuality and Reasoning Ability","date":"2025-03-04","arxiv_id":"2503.02343","repositories_listed":1,"syntology":null},{"url":"/paper/truth-knows-no-language-evaluating","slug":"truth-knows-no-language-evaluating","title":"Truth Knows No Language: Evaluating Truthfulness Beyond English","date":"2025-02-13","arxiv_id":"2502.09387","repositories_listed":1,"syntology":{"n":1,"n_ran":0,"n_constructed":0,"n_ran_checked":0,"n_instrument":0,"n_unverified":1,"n_honours":0,"n_violates":0,"n_no_contract":0,"n_pointer_only":0,"phrase":"0 ran · 1 unverified","sample_list":"/paper/truth-knows-no-language-evaluating#ran","syntology_url":"https://syntology.ai/paper/2502.09387","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2502.09387"}},"official":{"repos":["hitz-zentroa/truthfulqa-multi"],"state":"official: not harvested","n_ran":0,"n_constructed":0,"n_ran_no_instrument_failure":0,"n_unverified":0,"ran_from_kinds":[]}}},{"url":"/paper/multi-agent-reinforcement-learning-with-focal","slug":"multi-agent-reinforcement-learning-with-focal","title":"Multi-Agent Reinforcement Learning with Focal Diversity Optimization","date":"2025-02-06","arxiv_id":"2502.04492","repositories_listed":1,"syntology":{"n":1,"n_ran":1,"n_constructed":1,"n_ran_checked":1,"n_instrument":0,"n_unverified":0,"n_honours":0,"n_violates":0,"n_no_contract":1,"n_pointer_only":1,"phrase":"1 ran (of which 1 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 0 violated, 1 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified; the one sample that ran constructed an object rather than computing a result","sample_list":"/paper/multi-agent-reinforcement-learning-with-focal#ran","syntology_url":"https://syntology.ai/paper/2502.04492","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2502.04492"}},"official":{"repos":["sftekin/rl-focal"],"state":"official (archive's flag): 1 ran","n_ran":1,"n_constructed":1,"n_ran_no_instrument_failure":1,"n_unverified":0,"ran_from_kinds":["official"]}}},{"url":"/paper/chair-classifier-of-hallucination-as-improver","slug":"chair-classifier-of-hallucination-as-improver","title":"CHAIR -- Classifier of Hallucination as Improver","date":"2025-01-05","arxiv_id":"2501.02518","repositories_listed":1,"syntology":null},{"url":"/paper/whyphi-fine-tuning-phi-3-for-multiple-choice","slug":"whyphi-fine-tuning-phi-3-for-multiple-choice","title":"(WhyPHI) Fine-Tuning PHI-3 for Multiple-Choice Question Answering: Methodology, Results, and Challenges","date":"2025-01-03","arxiv_id":"2501.01588","repositories_listed":1,"syntology":null},{"url":"/paper/novo-norm-voting-off-hallucinations-with","slug":"novo-norm-voting-off-hallucinations-with","title":"NoVo: Norm Voting off Hallucinations with Attention Heads in Large Language Models","date":"2024-10-11","arxiv_id":"2410.08970","repositories_listed":1,"syntology":{"n":1,"n_ran":1,"n_constructed":0,"n_ran_checked":0,"n_instrument":1,"n_unverified":0,"n_honours":0,"n_violates":0,"n_no_contract":0,"n_pointer_only":1,"phrase":"1 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 1 where Syntology's instrument failed) · 0 unverified","sample_list":"/paper/novo-norm-voting-off-hallucinations-with#ran","syntology_url":"https://syntology.ai/paper/2410.08970","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2410.08970"}},"official":null}},{"url":"/paper/a-test-suite-of-prompt-injection-attacks-for","slug":"a-test-suite-of-prompt-injection-attacks-for","title":"A test suite of prompt injection attacks for LLM-based machine translation","date":"2024-10-07","arxiv_id":"2410.05047","repositories_listed":1,"syntology":null},{"url":"/paper/integrative-decoding-improve-factuality-via","slug":"integrative-decoding-improve-factuality-via","title":"Integrative Decoding: Improve Factuality via Implicit Self-consistency","date":"2024-10-02","arxiv_id":"2410.01556","repositories_listed":1,"syntology":{"n":3,"n_ran":1,"n_constructed":0,"n_ran_checked":0,"n_instrument":1,"n_unverified":2,"n_honours":0,"n_violates":0,"n_no_contract":0,"n_pointer_only":3,"phrase":"1 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 1 where Syntology's instrument failed) · 2 unverified","sample_list":"/paper/integrative-decoding-improve-factuality-via#ran","syntology_url":"https://syntology.ai/paper/2410.01556","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2410.01556"}},"official":{"repos":["yicheng98/integrativedecoding"],"state":"official (archive's flag): 1 ran","n_ran":1,"n_constructed":0,"n_ran_no_instrument_failure":0,"n_unverified":2,"ran_from_kinds":["official"]}}},{"url":"/paper/texttt-metabench-a-sparse-benchmark-to","slug":"texttt-metabench-a-sparse-benchmark-to","title":"$\\texttt{metabench}$ -- A Sparse Benchmark to Measure General Ability in Large Language Models","date":"2024-07-04","arxiv_id":"2407.12844","repositories_listed":1,"syntology":{"n":9,"n_ran":8,"n_constructed":0,"n_ran_checked":3,"n_instrument":5,"n_unverified":1,"n_honours":0,"n_violates":0,"n_no_contract":3,"n_pointer_only":9,"phrase":"8 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 5 where Syntology's instrument failed) · 1 unverified","sample_list":"/paper/texttt-metabench-a-sparse-benchmark-to#ran","syntology_url":"https://syntology.ai/paper/2407.12844","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2407.12844"}},"official":{"repos":["adkipnis/metabench"],"state":"official (archive's flag): 8 ran","n_ran":8,"n_constructed":0,"n_ran_no_instrument_failure":3,"n_unverified":1,"ran_from_kinds":["official"]}}},{"url":"/paper/varbench-robust-language-model-benchmarking","slug":"varbench-robust-language-model-benchmarking","title":"VarBench: Robust Language Model Benchmarking Through Dynamic Variable Perturbation","date":"2024-06-25","arxiv_id":"2406.17681","repositories_listed":1,"syntology":{"n":6,"n_ran":6,"n_constructed":0,"n_ran_checked":2,"n_instrument":4,"n_unverified":0,"n_honours":0,"n_violates":0,"n_no_contract":2,"n_pointer_only":6,"phrase":"6 ran (of which 0 constructed an object rather than computing a result; 2 with no instrument failure: 0 honoured, 0 violated, 2 with no contract checked; 4 where Syntology's instrument failed) · 0 unverified","sample_list":"/paper/varbench-robust-language-model-benchmarking#ran","syntology_url":"https://syntology.ai/paper/2406.17681","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2406.17681"}},"official":{"repos":["qbetterk/VarBench"],"state":"official (archive's flag): 6 ran","n_ran":6,"n_constructed":0,"n_ran_no_instrument_failure":2,"n_unverified":0,"ran_from_kinds":["official"]}}},{"url":"/paper/steering-without-side-effects-improving-post","slug":"steering-without-side-effects-improving-post","title":"Steering Without Side Effects: Improving Post-Deployment Control of Language Models","date":"2024-06-21","arxiv_id":"2406.15518","repositories_listed":1,"syntology":{"n":2,"n_ran":2,"n_constructed":0,"n_ran_checked":1,"n_instrument":1,"n_unverified":0,"n_honours":0,"n_violates":1,"n_no_contract":0,"n_pointer_only":2,"phrase":"2 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 1 violated, 0 with no contract checked; 1 where Syntology's instrument failed) · 0 unverified","sample_list":"/paper/steering-without-side-effects-improving-post#ran","syntology_url":"https://syntology.ai/paper/2406.15518","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2406.15518"}},"official":{"repos":["asacooperstickland/kl-then-steer"],"state":"official (archive's flag): 2 ran","n_ran":2,"n_constructed":0,"n_ran_no_instrument_failure":1,"n_unverified":0,"ran_from_kinds":["official"]}}},{"url":"/paper/enhancing-language-model-factuality-via","slug":"enhancing-language-model-factuality-via","title":"Enhancing Language Model Factuality via Activation-Based Confidence Calibration and Guided Decoding","date":"2024-06-19","arxiv_id":"2406.13230","repositories_listed":1,"syntology":null},{"url":"/paper/machine-unlearning-in-large-language-models","slug":"machine-unlearning-in-large-language-models","title":"Machine Unlearning in Large Language Models","date":"2024-05-24","arxiv_id":"2405.15152","repositories_listed":1,"syntology":null},{"url":"/paper/instruction-tuning-with-loss-over","slug":"instruction-tuning-with-loss-over","title":"Instruction Tuning With Loss Over Instructions","date":"2024-05-23","arxiv_id":"2405.14394","repositories_listed":1,"syntology":{"n":20,"n_ran":16,"n_constructed":0,"n_ran_checked":11,"n_instrument":5,"n_unverified":4,"n_honours":0,"n_violates":1,"n_no_contract":10,"n_pointer_only":20,"phrase":"16 ran (of which 0 constructed an object rather than computing a result; 11 with no instrument failure: 0 honoured, 1 violated, 10 with no contract checked; 5 where Syntology's instrument failed) · 4 unverified","sample_list":"/paper/instruction-tuning-with-loss-over#ran","syntology_url":"https://syntology.ai/paper/2405.14394","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2405.14394"}},"official":{"repos":["zhengxiangshi/instructionmodelling"],"state":"official (archive's flag): 16 ran","n_ran":16,"n_constructed":0,"n_ran_no_instrument_failure":11,"n_unverified":4,"ran_from_kinds":["official"]}}},{"url":"/paper/when-hindsight-is-not-20-20-testing-limits-on","slug":"when-hindsight-is-not-20-20-testing-limits-on","title":"When Hindsight is Not 20/20: Testing Limits on Reflective Thinking in Large Language Models","date":"2024-04-14","arxiv_id":"2404.09129","repositories_listed":1,"syntology":{"n":2,"n_ran":2,"n_constructed":0,"n_ran_checked":0,"n_instrument":2,"n_unverified":0,"n_honours":0,"n_violates":0,"n_no_contract":0,"n_pointer_only":2,"phrase":"2 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 2 where Syntology's instrument failed) · 0 unverified","sample_list":"/paper/when-hindsight-is-not-20-20-testing-limits-on#ran","syntology_url":"https://syntology.ai/paper/2404.09129","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2404.09129"}},"official":{"repos":["yanhong-lbh/llm-selfreflection-eval"],"state":"official (archive's flag): 2 ran","n_ran":2,"n_constructed":0,"n_ran_no_instrument_failure":0,"n_unverified":0,"ran_from_kinds":["official"]}}},{"url":"/paper/pollmgraph-unraveling-hallucinations-in-large","slug":"pollmgraph-unraveling-hallucinations-in-large","title":"PoLLMgraph: Unraveling Hallucinations in Large Language Models via State Transition Dynamics","date":"2024-04-06","arxiv_id":"2404.04722","repositories_listed":1,"syntology":{"n":7,"n_ran":6,"n_constructed":0,"n_ran_checked":3,"n_instrument":3,"n_unverified":1,"n_honours":0,"n_violates":1,"n_no_contract":2,"n_pointer_only":7,"phrase":"6 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 1 violated, 2 with no contract checked; 3 where Syntology's instrument failed) · 1 unverified","sample_list":"/paper/pollmgraph-unraveling-hallucinations-in-large#ran","syntology_url":"https://syntology.ai/paper/2404.04722","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2404.04722"}},"official":{"repos":["hitum-dev/pollmgraph"],"state":"official (archive's flag): 6 ran","n_ran":6,"n_constructed":0,"n_ran_no_instrument_failure":3,"n_unverified":1,"ran_from_kinds":["official"]}}},{"url":"/paper/nl-iti-optimizing-probing-and-intervention","slug":"nl-iti-optimizing-probing-and-intervention","title":"Non-Linear Inference Time Intervention: Improving LLM Truthfulness","date":"2024-03-27","arxiv_id":"2403.18680","repositories_listed":1,"syntology":null},{"url":"/paper/in-context-sharpness-as-alerts-an-inner","slug":"in-context-sharpness-as-alerts-an-inner","title":"In-Context Sharpness as Alerts: An Inner Representation Perspective for Hallucination Mitigation","date":"2024-03-03","arxiv_id":"2403.01548","repositories_listed":1,"syntology":{"n":4,"n_ran":2,"n_constructed":0,"n_ran_checked":0,"n_instrument":2,"n_unverified":2,"n_honours":0,"n_violates":0,"n_no_contract":0,"n_pointer_only":4,"phrase":"2 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 2 where Syntology's instrument failed) · 2 unverified","sample_list":"/paper/in-context-sharpness-as-alerts-an-inner#ran","syntology_url":"https://syntology.ai/paper/2403.01548","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2403.01548"}},"official":{"repos":["hkust-nlp/activation_decoding"],"state":"official (archive's flag): 2 ran","n_ran":2,"n_constructed":0,"n_ran_no_instrument_failure":0,"n_unverified":1,"ran_from_kinds":["official"]}}},{"url":"/paper/truthx-alleviating-hallucinations-by-editing","slug":"truthx-alleviating-hallucinations-by-editing","title":"TruthX: Alleviating Hallucinations by Editing Large Language Models in Truthful Space","date":"2024-02-27","arxiv_id":"2402.17811","repositories_listed":1,"syntology":{"n":3,"n_ran":2,"n_constructed":2,"n_ran_checked":2,"n_instrument":0,"n_unverified":1,"n_honours":0,"n_violates":0,"n_no_contract":2,"n_pointer_only":3,"phrase":"2 ran (of which 2 constructed an object rather than computing a result; 2 with no instrument failure: 0 honoured, 0 violated, 2 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified; every one of the 2 samples that ran constructed an object rather than computing a result","sample_list":"/paper/truthx-alleviating-hallucinations-by-editing#ran","syntology_url":"https://syntology.ai/paper/2402.17811","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2402.17811"}},"official":{"repos":["ictnlp/truthx"],"state":"official (archive's flag): 2 ran","n_ran":2,"n_constructed":2,"n_ran_no_instrument_failure":2,"n_unverified":1,"ran_from_kinds":["official"]}}},{"url":"/paper/sage-evaluating-moral-consistency-in-large","slug":"sage-evaluating-moral-consistency-in-large","title":"SaGE: Evaluating Moral Consistency in Large Language Models","date":"2024-02-21","arxiv_id":"2402.13709","repositories_listed":1,"syntology":{"n":1,"n_ran":1,"n_constructed":0,"n_ran_checked":0,"n_instrument":1,"n_unverified":0,"n_honours":0,"n_violates":0,"n_no_contract":0,"n_pointer_only":1,"phrase":"1 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 1 where Syntology's instrument failed) · 0 unverified","sample_list":"/paper/sage-evaluating-moral-consistency-in-large#ran","syntology_url":"https://syntology.ai/paper/2402.13709","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2402.13709"}},"official":{"repos":["vnnm404/SaGE"],"state":"official (archive's flag): 1 ran","n_ran":1,"n_constructed":0,"n_ran_no_instrument_failure":0,"n_unverified":0,"ran_from_kinds":["official"]}}},{"url":"/paper/truth-forest-toward-multi-scale-truthfulness","slug":"truth-forest-toward-multi-scale-truthfulness","title":"Truth Forest: Toward Multi-Scale Truthfulness in Large Language Models through Intervention without Tuning","date":"2023-12-29","arxiv_id":"2312.17484","repositories_listed":1,"syntology":null},{"url":"/paper/instruction-tuning-with-human-curriculum","slug":"instruction-tuning-with-human-curriculum","title":"Instruction Tuning with Human Curriculum","date":"2023-10-14","arxiv_id":"2310.09518","repositories_listed":1,"syntology":null},{"url":"/paper/tool-augmented-reward-modeling","slug":"tool-augmented-reward-modeling","title":"Tool-Augmented Reward Modeling","date":"2023-10-02","arxiv_id":"2310.01045","repositories_listed":1,"syntology":{"n":8,"n_ran":8,"n_constructed":0,"n_ran_checked":7,"n_instrument":1,"n_unverified":0,"n_honours":0,"n_violates":0,"n_no_contract":7,"n_pointer_only":0,"phrase":"8 ran (of which 0 constructed an object rather than computing a result; 7 with no instrument failure: 0 honoured, 0 violated, 7 with no contract checked; 1 where Syntology's instrument failed) · 0 unverified","sample_list":"/paper/tool-augmented-reward-modeling#ran","syntology_url":"https://syntology.ai/paper/2310.01045","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2310.01045"}},"official":{"repos":["ernie-research/Tool-Augmented-Reward-Model"],"state":"official (archive's flag): 8 ran","n_ran":8,"n_constructed":0,"n_ran_no_instrument_failure":7,"n_unverified":0,"ran_from_kinds":["official"]}}},{"url":"/paper/rain-your-language-models-can-align","slug":"rain-your-language-models-can-align","title":"RAIN: Your Language Models Can Align Themselves without Finetuning","date":"2023-09-13","arxiv_id":"2309.07124","repositories_listed":1,"syntology":{"n":1,"n_ran":1,"n_constructed":0,"n_ran_checked":1,"n_instrument":0,"n_unverified":0,"n_honours":1,"n_violates":0,"n_no_contract":0,"n_pointer_only":0,"phrase":"1 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 1 honoured, 0 violated, 0 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified","sample_list":"/paper/rain-your-language-models-can-align#ran","syntology_url":"https://syntology.ai/paper/2309.07124","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2309.07124"}},"official":{"repos":["SafeAILab/RAIN"],"state":"official (archive's flag): 1 ran","n_ran":1,"n_constructed":0,"n_ran_no_instrument_failure":1,"n_unverified":0,"ran_from_kinds":["official"]}}},{"url":"/paper/sight-beyond-text-multi-modal-training","slug":"sight-beyond-text-multi-modal-training","title":"Sight Beyond Text: Multi-Modal Training Enhances LLMs in Truthfulness and Ethics","date":"2023-09-13","arxiv_id":"2309.07120","repositories_listed":1,"syntology":{"n":6,"n_ran":4,"n_constructed":0,"n_ran_checked":3,"n_instrument":1,"n_unverified":2,"n_honours":1,"n_violates":1,"n_no_contract":1,"n_pointer_only":2,"phrase":"4 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 1 honoured, 1 violated, 1 with no contract checked; 1 where Syntology's instrument failed) · 2 unverified","sample_list":"/paper/sight-beyond-text-multi-modal-training#ran","syntology_url":"https://syntology.ai/paper/2309.07120","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2309.07120"}},"official":{"repos":["ucsc-vlaa/sight-beyond-text"],"state":"official (archive's flag): 4 ran","n_ran":4,"n_constructed":0,"n_ran_no_instrument_failure":3,"n_unverified":2,"ran_from_kinds":["official"]}}},{"url":"/paper/inference-time-intervention-eliciting","slug":"inference-time-intervention-eliciting","title":"Inference-Time Intervention: Eliciting Truthful Answers from a Language Model","date":"2023-06-06","arxiv_id":"2306.03341","repositories_listed":1,"syntology":{"n":1,"n_ran":1,"n_constructed":1,"n_ran_checked":1,"n_instrument":0,"n_unverified":0,"n_honours":0,"n_violates":0,"n_no_contract":1,"n_pointer_only":0,"phrase":"1 ran (of which 1 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 0 violated, 1 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified; the one sample that ran constructed an object rather than computing a result","sample_list":"/paper/inference-time-intervention-eliciting#ran","syntology_url":"https://syntology.ai/paper/2306.03341","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2306.03341"}},"official":{"repos":["likenneth/honest_llama"],"state":"official (archive's flag): 1 ran","n_ran":1,"n_constructed":1,"n_ran_no_instrument_failure":1,"n_unverified":0,"ran_from_kinds":["official"]}}},{"url":"/paper/just-ask-for-calibration-strategies-for","slug":"just-ask-for-calibration-strategies-for","title":"Just Ask for Calibration: Strategies for Eliciting Calibrated Confidence Scores from Language Models Fine-Tuned with Human Feedback","date":"2023-05-24","arxiv_id":"2305.14975","repositories_listed":1,"syntology":null},{"url":"/paper/measuring-reliability-of-large-language","slug":"measuring-reliability-of-large-language","title":"Measuring Reliability of Large Language Models through Semantic Consistency","date":"2022-11-10","arxiv_id":"2211.05853","repositories_listed":1,"syntology":null},{"url":null,"slug":"model-unlearning-via-sparse-autoencoder","title":"Model Unlearning via Sparse Autoencoder Subspace Guided Projections","date":"2025-05-30","arxiv_id":"2505.24428","repositories_listed":0,"syntology":null},{"url":null,"slug":"shadows-in-the-attention-contextual","title":"Shadows in the Attention: Contextual Perturbation and Representation Drift in the Dynamics of Hallucination in LLMs","date":"2025-05-22","arxiv_id":"2505.16894","repositories_listed":0,"syntology":null},{"url":null,"slug":"elastic-weight-consolidation-for-full","title":"Elastic Weight Consolidation for Full-Parameter Continual Pre-Training of Gemma2","date":"2025-05-09","arxiv_id":"2505.05946","repositories_listed":0,"syntology":null},{"url":null,"slug":"dynamax-dynamic-computing-for-transformers","title":"DYNAMAX: Dynamic computing for Transformers and Mamba based architectures","date":"2025-04-29","arxiv_id":"2504.20922","repositories_listed":0,"syntology":null},{"url":null,"slug":"efficient-map-estimation-of-llm-judgment","title":"Efficient MAP Estimation of LLM Judgment Performance with Prior Transfer","date":"2025-04-17","arxiv_id":"2504.12589","repositories_listed":0,"syntology":null},{"url":null,"slug":"sample-don-t-search-rethinking-test-time","title":"Sample, Don't Search: Rethinking Test-Time Alignment for Language Models","date":"2025-04-04","arxiv_id":"2504.03790","repositories_listed":0,"syntology":null},{"url":null,"slug":"sustainable-llm-inference-for-edge-ai","title":"Sustainable LLM Inference for Edge AI: Evaluating Quantized LLMs for Energy Efficiency, Output Accuracy, and Inference Latency","date":"2025-04-04","arxiv_id":"2504.03360","repositories_listed":0,"syntology":null},{"url":null,"slug":"more-is-less-the-pitfalls-of-multi-model","title":"More is Less: The Pitfalls of Multi-Model Synthetic Preference Data in DPO Safety Alignment","date":"2025-04-03","arxiv_id":"2504.02193","repositories_listed":0,"syntology":null},{"url":null,"slug":"when-persuasion-overrides-truth-in-multi","title":"When Persuasion Overrides Truth in Multi-Agent LLM Debates: Introducing a Confidence-Weighted Persuasion Override Rate (CW-POR)","date":"2025-04-01","arxiv_id":"2504.00374","repositories_listed":0,"syntology":null},{"url":null,"slug":"obliviate-efficient-unmemorization-for","title":"Obliviate: Efficient Unmemorization for Protecting Intellectual Property in Large Language Models","date":"2025-02-20","arxiv_id":"2502.15010","repositories_listed":0,"syntology":null},{"url":null,"slug":"cost-saving-llm-cascades-with-early","title":"Cost-Saving LLM Cascades with Early Abstention","date":"2025-02-13","arxiv_id":"2502.09054","repositories_listed":0,"syntology":null},{"url":null,"slug":"selective-self-to-supervised-fine-tuning-for","title":"Selective Self-to-Supervised Fine-Tuning for Generalization in Large Language Models","date":"2025-02-12","arxiv_id":"2502.08130","repositories_listed":0,"syntology":null},{"url":null,"slug":"truthflow-truthful-llm-generation-via","title":"TruthFlow: Truthful LLM Generation via Representation Flow Correction","date":"2025-02-06","arxiv_id":"2502.04556","repositories_listed":0,"syntology":null},{"url":null,"slug":"monty-hall-and-optimized-conformal-prediction","title":"Monty Hall and Optimized Conformal Prediction to Improve Decision-Making with LLMs","date":"2024-12-31","arxiv_id":"2501.00555","repositories_listed":0,"syntology":null},{"url":null,"slug":"mitigating-adversarial-attacks-in-llms","title":"Mitigating Adversarial Attacks in LLMs through Defensive Suffix Generation","date":"2024-12-18","arxiv_id":"2412.13705","repositories_listed":0,"syntology":null},{"url":null,"slug":"uhura-a-benchmark-for-evaluating-scientific","title":"Uhura: A Benchmark for Evaluating Scientific Question Answering and Truthfulness in Low-Resource African Languages","date":"2024-12-01","arxiv_id":"2412.00948","repositories_listed":0,"syntology":null},{"url":null,"slug":"layer-importance-and-hallucination-analysis","title":"Layer Importance and Hallucination Analysis in Large Language Models via Enhanced Activation Variance-Sparsity","date":"2024-11-15","arxiv_id":"2411.10069","repositories_listed":0,"syntology":null},{"url":null,"slug":"maintaining-informative-coherence-migrating","title":"Maintaining Informative Coherence: Migrating Hallucinations in Large Language Models via Absorbing Markov Chains","date":"2024-10-27","arxiv_id":"2410.20340","repositories_listed":0,"syntology":null},{"url":null,"slug":"a-debate-driven-experiment-on-llm","title":"A Debate-Driven Experiment on LLM Hallucinations and Accuracy","date":"2024-10-25","arxiv_id":"2410.19485","repositories_listed":0,"syntology":null},{"url":null,"slug":"evaluating-consistencies-in-llm-responses","title":"Evaluating Consistencies in LLM responses through a Semantic Clustering of Question Answering","date":"2024-10-20","arxiv_id":"2410.15440","repositories_listed":0,"syntology":null},{"url":null,"slug":"iter-ahmcl-alleviate-hallucination-for-large","title":"Iter-AHMCL: Alleviate Hallucination for Large Language Model via Iterative Model-level Contrastive Learning","date":"2024-10-16","arxiv_id":"2410.12130","repositories_listed":0,"syntology":null},{"url":null,"slug":"skillaggregation-reference-free-llm-dependent","title":"SkillAggregation: Reference-free LLM-Dependent Aggregation","date":"2024-10-14","arxiv_id":"2410.10215","repositories_listed":0,"syntology":null},{"url":null,"slug":"benchmark-inflation-revealing-llm-performance","title":"Benchmark Inflation: Revealing LLM Performance Gaps Using Retro-Holdouts","date":"2024-10-11","arxiv_id":"2410.09247","repositories_listed":0,"syntology":null},{"url":null,"slug":"towards-cross-lingual-llm-evaluation-for","title":"Towards Multilingual LLM Evaluation for European Languages","date":"2024-10-11","arxiv_id":"2410.08928","repositories_listed":0,"syntology":null},{"url":null,"slug":"efficiently-deploying-llms-with-controlled","title":"Efficiently Deploying LLMs with Controlled Risk","date":"2024-10-03","arxiv_id":"2410.02173","repositories_listed":0,"syntology":null},{"url":null,"slug":"progress-report-towards-european-llms","title":"Teuken-7B-Base & Teuken-7B-Instruct: Towards European LLMs","date":"2024-09-30","arxiv_id":"2410.03730","repositories_listed":0,"syntology":null},{"url":null,"slug":"selective-self-rehearsal-a-fine-tuning","title":"Selective Self-Rehearsal: A Fine-Tuning Approach to Improve Generalization in Large Language Models","date":"2024-09-07","arxiv_id":"2409.04787","repositories_listed":0,"syntology":null},{"url":null,"slug":"lower-layer-matters-alleviating-hallucination","title":"Lower Layer Matters: Alleviating Hallucination via Multi-Layer Fusion Contrastive Decoding with Truthfulness Refocused","date":"2024-08-16","arxiv_id":"2408.08769","repositories_listed":0,"syntology":null},{"url":null,"slug":"lokilm-technical-report","title":"LokiLM: Technical Report","date":"2024-07-10","arxiv_id":"2407.07370","repositories_listed":0,"syntology":null},{"url":null,"slug":"multi-reference-preference-optimization-for","title":"Multi-Reference Preference Optimization for Large Language Models","date":"2024-05-26","arxiv_id":"2405.16388","repositories_listed":0,"syntology":null},{"url":null,"slug":"harmonic-llms-are-trustworthy","title":"Harmonic LLMs are Trustworthy","date":"2024-04-30","arxiv_id":"2404.19708","repositories_listed":0,"syntology":null},{"url":null,"slug":"regressive-side-effects-of-training-language","title":"Student Data Paradox and Curious Case of Single Student-Tutor Model: Regressive Side Effects of Training LLMs for Personalized Learning","date":"2024-04-23","arxiv_id":"2404.15156","repositories_listed":0,"syntology":null},{"url":null,"slug":"probelm-plausibility-ranking-evaluation-for","title":"PRobELM: Plausibility Ranking Evaluation for Language Models","date":"2024-04-04","arxiv_id":"2404.03818","repositories_listed":0,"syntology":null},{"url":null,"slug":"developing-a-framework-for-auditing-large","title":"LLMAuditor: A Framework for Auditing Large Language Models Using Human-in-the-Loop","date":"2024-02-14","arxiv_id":"2402.09346","repositories_listed":0,"syntology":null},{"url":null,"slug":"self-alignment-for-factuality-mitigating","title":"Self-Alignment for Factuality: Mitigating Hallucinations in LLMs via Self-Evaluation","date":"2024-02-14","arxiv_id":"2402.09267","repositories_listed":0,"syntology":null},{"url":null,"slug":"grath-gradual-self-truthifying-for-large","title":"GRATH: Gradual Self-Truthifying for Large Language Models","date":"2024-01-22","arxiv_id":"2401.12292","repositories_listed":0,"syntology":null},{"url":null,"slug":"reducing-llm-hallucinations-using-epistemic","title":"Reducing LLM Hallucinations using Epistemic Neural Networks","date":"2023-12-25","arxiv_id":"2312.15576","repositories_listed":0,"syntology":null},{"url":null,"slug":"self-evaluation-improves-selective-generation","title":"Self-Evaluation Improves Selective Generation in Large Language Models","date":"2023-12-14","arxiv_id":"2312.09300","repositories_listed":0,"syntology":null},{"url":null,"slug":"uncertainty-aware-language-modeling-for","title":"Uncertainty-aware Language Modeling for Selective Question Answering","date":"2023-11-26","arxiv_id":"2311.15451","repositories_listed":0,"syntology":null},{"url":null,"slug":"investigating-data-contamination-in-modern","title":"Investigating Data Contamination in Modern Benchmarks for Large Language Models","date":"2023-11-16","arxiv_id":"2311.09783","repositories_listed":0,"syntology":null},{"url":null,"slug":"on-the-truthfulness-of-surprisingly-likely","title":"On The Truthfulness of 'Surprisingly Likely' Responses of Large Language Models","date":"2023-11-13","arxiv_id":"2311.07692","repositories_listed":0,"syntology":null},{"url":null,"slug":"semantic-consistency-for-assuring-reliability","title":"Semantic Consistency for Assuring Reliability of Large Language Models","date":"2023-08-17","arxiv_id":"2308.09138","repositories_listed":0,"syntology":null},{"url":null,"slug":"teaching-language-models-to-support-answers","title":"Teaching language models to support answers with verified quotes","date":"2022-03-21","arxiv_id":"2203.11147","repositories_listed":0,"syntology":null}],"syntology_records":22,"syntology_note":"a paper without a record is not a recorded non-run: it may lack an arXiv id or simply be absent from the graph layer"},"key_notes":{"n_ran_checked":"legacy name, kept unchanged so existing readers do not break: it counts the samples that ran with no instrument failure (honoured, violated, and ran with no contract checked); it does not mean a contract was checked, and the pages print it as 'K with no instrument failure', not 'K checked'","n_constructed":"a sub-count of the samples that ran, never subtracted from them and never a failure: an executed sample whose run returned an instance of its own class (fixture_out_type equals the entry name): the run built an object and did not compute a result (Syntology's RAN record, counts.constructed)"},"since_archive_judged":{"papers":0,"papers_placed":0,"judged_by":{"model":"a model","models_as_recorded":[],"from":"lb_rows_run_verified.jsonl: each record's `model` field (the extractor's call that wrote decisions[])"},"note":"a model's judgement from the paper's own tables that they report on this leaderboard, not a result: no value from a paper that was not placed is printed or carried here; papers counts each paper once over these leaderboards","backfill":"8,886 of 18,637 newer papers checked so far (733 more had nothing to place; 41 could not be read)","backfill_complete":false,"leaderboards":[]},"papers_where_code_ran":20,"papers_where_code_ran_first_page":"/task/truthfulqa/papers/ran/1","papers_where_code_ran_caption":"We ran code from the paper's repository; we did not run it on this task or check it against the task's benchmarks.","pages":[{"order":"ran","page":1,"of":1,"url":"/task/truthfulqa/papers/ran/1","json":"/task/truthfulqa/papers/ran/1.json","rows":20}],"description_links":{"kept":0,"unwrapped_to_text":0,"bare_urls_linked":0,"relative_images_dropped":0,"rule":"internal links are kept only when the target slug exists in the catalog"},"syntology":{"read_at":"2026-10-05T00:36:11+00:00","claim":"Per-sample execution status on synthesized fixtures ('ran N of M samples'); not a correctness claim and not a ranking signal.","status_vocabulary":{"ran_honours":"ran, honoured the contract we drafted","ran_violates":"ran, violated the contract we drafted","ran_draft_wrong":"ran; our contract draft was wrong, not the code","ran_fixture":"ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)","ran":"ran on a synthesized input","unverified":"unverified (harvested, no recorded run)"},"row_route":"each syntology object on a benchmark's first row and on a paper carries sample_list (this site's paper page, #ran), syntology_url (syntology.ai's page for the paper, only when that route probed live) and mcp (the agent call that lists the paper's harvested samples; not a verdict on the paper)","developers":"https://syntology.ai/developers"},"not_shown":{"libraries":"the archive has no per-task library table","trend_sparklines":"the Trend column of the benchmarks table was a rendered image; it is not in the archive","social_and_latest_sorts":"stars and social signals are not in the archive"},"record_sha256":"6d28647d4eeb38b1cd19c1e206d239fde5dccd7a9311e33ec0727b99b8ea9bbf","record_changed_at":"2026-10-05","record_changed_at_basis":"changed"}