{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:DRCNPBOL4LFP4KPSBUUH2GHM3S","short_pith_number":"pith:DRCNPBOL","schema_version":"1.0","canonical_sha256":"1c44d785cbe2cafe29f20d287d18ecdca6420aaac85dabee8dacae5215f2c33d","source":{"kind":"arxiv","id":"2605.12922","version":1},"attestation_state":"computed","paper":{"title":"When Attention Closes: How LLMs Lose the Thread in Multi-Turn Interaction","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"LLMs lose track of instructions in multi-turn chats when attention to goal tokens fades, though residual streams may still encode the needed information.","cross_cats":["cs.CL"],"primary_cat":"cs.AI","authors_text":"Dilek Hakkani-T\\\"ur, Joseph Hsieh, Seunghyun Yoon, Trung Bui, Vardhan Dongre, Viet Dac Lai","submitted_at":"2026-05-13T02:58:18Z","abstract_excerpt":"Large language models can follow complex instructions in a single turn, yet over long multi-turn interactions they often lose the thread of instructions, persona, and rules. This degradation has been measured behaviorally but not mechanistically explained. We propose a channel-transition account: goal-defining tokens become less accessible through attention, while goal-related information may persist in residual representations. We introduce the Goal Accessibility Ratio (GAR), measuring attention from generated tokens to task-defining goal tokens, and combine it with sliding-window ablations a"},"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":true,"formal_links_present":true},"canonical_record":{"source":{"id":"2605.12922","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.AI","submitted_at":"2026-05-13T02:58:18Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"2bbf3223d3da58d082e2aedb69c8f93e0a5c66f6ef707e03a30352b81963495e","abstract_canon_sha256":"6c8c5dbb6f4bc4438f88cdee9f6f6be59c17805b60a5aa72fe5788b4dc225d20"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T03:09:10.200250Z","signature_b64":"PrFZW5XzIApoekeR09R3gm0+GuAhBUrzietI90YTozguSIj0PniLyGpYPPRCMzQGTCLqhOdFEbo3DMofisdsBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1c44d785cbe2cafe29f20d287d18ecdca6420aaac85dabee8dacae5215f2c33d","last_reissued_at":"2026-05-18T03:09:10.199507Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T03:09:10.199507Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"When Attention Closes: How LLMs Lose the Thread in Multi-Turn Interaction","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"LLMs lose track of instructions in multi-turn chats when attention to goal tokens fades, though residual streams may still encode the needed information.","cross_cats":["cs.CL"],"primary_cat":"cs.AI","authors_text":"Dilek Hakkani-T\\\"ur, Joseph Hsieh, Seunghyun Yoon, Trung Bui, Vardhan Dongre, Viet Dac Lai","submitted_at":"2026-05-13T02:58:18Z","abstract_excerpt":"Large language models can follow complex instructions in a single turn, yet over long multi-turn interactions they often lose the thread of instructions, persona, and rules. This degradation has been measured behaviorally but not mechanistically explained. We propose a channel-transition account: goal-defining tokens become less accessible through attention, while goal-related information may persist in residual representations. We introduce the Goal Accessibility Ratio (GAR), measuring attention from generated tokens to task-defining goal tokens, and combine it with sliding-window ablations a"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Across architectures and model scales, the gap between attention loss and residual decodability predicts whether goal-conditioned behavior survives channel closure.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the linear probes on residual streams recover causally relevant goal information rather than spurious correlations, and that the sliding-window ablations isolate the attention channel without confounding other mechanisms.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Attention to goal tokens declines in multi-turn LLM interactions while residual representations often retain decodable goal information, and the gap between these predicts whether goal-conditioned behavior survives.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"LLMs lose track of instructions in multi-turn chats when attention to goal tokens fades, though residual streams may still encode the needed information.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"66c60cca120bc073a4b1d313963698529129e4d39ab4a99e19780123d969800d"},"source":{"id":"2605.12922","kind":"arxiv","version":1},"verdict":{"id":"dbd61569-cdfb-4f4d-81c5-ff1b18d7e154","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-14T20:16:15.218921Z","strongest_claim":"Across architectures and model scales, the gap between attention loss and residual decodability predicts whether goal-conditioned behavior survives channel closure.","one_line_summary":"Attention to goal tokens declines in multi-turn LLM interactions while residual representations often retain decodable goal information, and the gap between these predicts whether goal-conditioned behavior survives.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the linear probes on residual streams recover causally relevant goal information rather than spurious correlations, and that the sliding-window ablations isolate the attention channel without confounding other mechanisms.","pith_extraction_headline":"LLMs lose track of instructions in multi-turn chats when attention to goal tokens fades, though residual streams may still encode the needed information."},"references":{"count":39,"sample":[{"doi":"","year":null,"title":"LLMs Get Lost In Multi-Turn Conversation","work_id":"01fa2ab4-4b40-4e8a-9487-2de7db19cfe2","ref_index":1,"cited_arxiv_id":"2505.06120","is_internal_anchor":true},{"doi":"","year":null,"title":"Multi-if: Benchmarking llms on multi-turn and multilingual instructions following","work_id":"2e46fb28-13e3-4c54-88b5-a2d47f2a064b","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"Transactions of the association for computational linguistics , volume=","work_id":"35c18954-8a0c-4935-b39c-5981194920ba","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2024,"title":"Roformer: Enhanced transformer with rotary position embedding , author=. Neurocomputing , volume=. 2024 , publisher=","work_id":"0cfd4ae7-837c-4c4f-bf6d-afc10f5a8ed1","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) , pages=","work_id":"756542c0-c2f6-4557-9f1f-50c1be6099a8","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":39,"snapshot_sha256":"e9bd9f0947179197d22903ff1caab17d502cfdd23f9c82b831f5c672746b816f","internal_anchors":17},"formal_canon":{"evidence_count":2,"snapshot_sha256":"c864106dc6b25539c1110e1faf9125625aec269b7b8ef15cc8dbc800ea61e563"},"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":"2605.12922","created_at":"2026-05-18T03:09:10.199632+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.12922v1","created_at":"2026-05-18T03:09:10.199632+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.12922","created_at":"2026-05-18T03:09:10.199632+00:00"},{"alias_kind":"pith_short_12","alias_value":"DRCNPBOL4LFP","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_16","alias_value":"DRCNPBOL4LFP4KPS","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_8","alias_value":"DRCNPBOL","created_at":"2026-05-18T12:33:37.589309+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":2,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/DRCNPBOL4LFP4KPSBUUH2GHM3S","json":"https://pith.science/pith/DRCNPBOL4LFP4KPSBUUH2GHM3S.json","graph_json":"https://pith.science/api/pith-number/DRCNPBOL4LFP4KPSBUUH2GHM3S/graph.json","events_json":"https://pith.science/api/pith-number/DRCNPBOL4LFP4KPSBUUH2GHM3S/events.json","paper":"https://pith.science/paper/DRCNPBOL"},"agent_actions":{"view_html":"https://pith.science/pith/DRCNPBOL4LFP4KPSBUUH2GHM3S","download_json":"https://pith.science/pith/DRCNPBOL4LFP4KPSBUUH2GHM3S.json","view_paper":"https://pith.science/paper/DRCNPBOL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.12922&json=true","fetch_graph":"https://pith.science/api/pith-number/DRCNPBOL4LFP4KPSBUUH2GHM3S/graph.json","fetch_events":"https://pith.science/api/pith-number/DRCNPBOL4LFP4KPSBUUH2GHM3S/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/DRCNPBOL4LFP4KPSBUUH2GHM3S/action/timestamp_anchor","attest_storage":"https://pith.science/pith/DRCNPBOL4LFP4KPSBUUH2GHM3S/action/storage_attestation","attest_author":"https://pith.science/pith/DRCNPBOL4LFP4KPSBUUH2GHM3S/action/author_attestation","sign_citation":"https://pith.science/pith/DRCNPBOL4LFP4KPSBUUH2GHM3S/action/citation_signature","submit_replication":"https://pith.science/pith/DRCNPBOL4LFP4KPSBUUH2GHM3S/action/replication_record"}},"created_at":"2026-05-18T03:09:10.199632+00:00","updated_at":"2026-05-18T03:09:10.199632+00:00"}