{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:GK6P2W4XCM5UHLKYH2LFQ3DLEM","short_pith_number":"pith:GK6P2W4X","schema_version":"1.0","canonical_sha256":"32bcfd5b97133b43ad583e96586c6b233933a9e829c6b117cd0bcfd54668bf70","source":{"kind":"arxiv","id":"2406.11050","version":2},"attestation_state":"computed","paper":{"title":"A Peek into Token Bias: Large Language Models Are Not Yet Genuine Reasoners","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Bowen Jiang, Camillo J. Taylor, Dan Roth, Tanwi Mallick, Weijie J. Su, Xiaomeng Wang, Yangxinyu Xie, Zhuoqun Hao","submitted_at":"2024-06-16T19:22:53Z","abstract_excerpt":"This study introduces a hypothesis-testing framework to assess whether large language models (LLMs) possess genuine reasoning abilities or primarily depend on token bias. We go beyond evaluating LLMs on accuracy; rather, we aim to investigate their token bias in solving logical reasoning tasks. Specifically, we develop carefully controlled synthetic datasets, featuring conjunction fallacy and syllogistic problems. Our framework outlines a list of hypotheses where token biases are readily identifiable, with all null hypotheses assuming genuine reasoning capabilities of LLMs. The findings in thi"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2406.11050","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-06-16T19:22:53Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"b00de5e89a550c0dd27f404c97ef091774b3c0a47b4464f66fdc6e1f73ac8b13","abstract_canon_sha256":"36a6c3e4adc943808aeec2e9851fe23b224fe80a12c173796d872f18538dcffe"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:15:33.699348Z","signature_b64":"LOa10peN5UplwpMnQSotZr9bX7Ljkr2O7DDCKCGTcSz0zBnzVtq7VMTFaYO6VrwvKuOTrfGPbJHetaji/w7kAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"32bcfd5b97133b43ad583e96586c6b233933a9e829c6b117cd0bcfd54668bf70","last_reissued_at":"2026-07-05T09:15:33.698862Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:15:33.698862Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"A Peek into Token Bias: Large Language Models Are Not Yet Genuine Reasoners","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Bowen Jiang, Camillo J. Taylor, Dan Roth, Tanwi Mallick, Weijie J. Su, Xiaomeng Wang, Yangxinyu Xie, Zhuoqun Hao","submitted_at":"2024-06-16T19:22:53Z","abstract_excerpt":"This study introduces a hypothesis-testing framework to assess whether large language models (LLMs) possess genuine reasoning abilities or primarily depend on token bias. We go beyond evaluating LLMs on accuracy; rather, we aim to investigate their token bias in solving logical reasoning tasks. Specifically, we develop carefully controlled synthetic datasets, featuring conjunction fallacy and syllogistic problems. Our framework outlines a list of hypotheses where token biases are readily identifiable, with all null hypotheses assuming genuine reasoning capabilities of LLMs. The findings in thi"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.11050","kind":"arxiv","version":2},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2406.11050/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2406.11050","created_at":"2026-07-05T09:15:33.698915+00:00"},{"alias_kind":"arxiv_version","alias_value":"2406.11050v2","created_at":"2026-07-05T09:15:33.698915+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.11050","created_at":"2026-07-05T09:15:33.698915+00:00"},{"alias_kind":"pith_short_12","alias_value":"GK6P2W4XCM5U","created_at":"2026-07-05T09:15:33.698915+00:00"},{"alias_kind":"pith_short_16","alias_value":"GK6P2W4XCM5UHLKY","created_at":"2026-07-05T09:15:33.698915+00:00"},{"alias_kind":"pith_short_8","alias_value":"GK6P2W4X","created_at":"2026-07-05T09:15:33.698915+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":9,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.07515","citing_title":"How reliable are LLMs when it comes to playing dice?","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2605.22714","citing_title":"AMEL: Accumulated Message Effects on LLM Judgments","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2503.18018","citing_title":"Lost in Cultural Translation: Do LLMs Struggle with Math Across Cultural Contexts?","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2605.22714","citing_title":"AMEL: Accumulated Message Effects on LLM Judgments","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2410.05229","citing_title":"GSM-Symbolic: Understanding the Limitations of Mathematical Reasoning in Large Language Models","ref_index":75,"is_internal_anchor":false},{"citing_arxiv_id":"2605.00369","citing_title":"InvEvolve: Evolving White-Box Inventory Policies via Large Language Models with Performance Guarantees","ref_index":156,"is_internal_anchor":false},{"citing_arxiv_id":"2412.05579","citing_title":"LLMs-as-Judges: A Comprehensive Survey on LLM-based Evaluation Methods","ref_index":102,"is_internal_anchor":false},{"citing_arxiv_id":"2605.00369","citing_title":"InvEvolve: Evolving White-Box Inventory Policies via Large Language Models with Performance Guarantees","ref_index":156,"is_internal_anchor":false},{"citing_arxiv_id":"2604.21632","citing_title":"To See the Unseen: on the Generalization Ability of Transformers in Symbolic Reasoning","ref_index":7,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/GK6P2W4XCM5UHLKYH2LFQ3DLEM","json":"https://pith.science/pith/GK6P2W4XCM5UHLKYH2LFQ3DLEM.json","graph_json":"https://pith.science/api/pith-number/GK6P2W4XCM5UHLKYH2LFQ3DLEM/graph.json","events_json":"https://pith.science/api/pith-number/GK6P2W4XCM5UHLKYH2LFQ3DLEM/events.json","paper":"https://pith.science/paper/GK6P2W4X"},"agent_actions":{"view_html":"https://pith.science/pith/GK6P2W4XCM5UHLKYH2LFQ3DLEM","download_json":"https://pith.science/pith/GK6P2W4XCM5UHLKYH2LFQ3DLEM.json","view_paper":"https://pith.science/paper/GK6P2W4X","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2406.11050&json=true","fetch_graph":"https://pith.science/api/pith-number/GK6P2W4XCM5UHLKYH2LFQ3DLEM/graph.json","fetch_events":"https://pith.science/api/pith-number/GK6P2W4XCM5UHLKYH2LFQ3DLEM/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/GK6P2W4XCM5UHLKYH2LFQ3DLEM/action/timestamp_anchor","attest_storage":"https://pith.science/pith/GK6P2W4XCM5UHLKYH2LFQ3DLEM/action/storage_attestation","attest_author":"https://pith.science/pith/GK6P2W4XCM5UHLKYH2LFQ3DLEM/action/author_attestation","sign_citation":"https://pith.science/pith/GK6P2W4XCM5UHLKYH2LFQ3DLEM/action/citation_signature","submit_replication":"https://pith.science/pith/GK6P2W4XCM5UHLKYH2LFQ3DLEM/action/replication_record"}},"created_at":"2026-07-05T09:15:33.698915+00:00","updated_at":"2026-07-05T09:15:33.698915+00:00"}