{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:FGRNTKQRG2FHKGWQSHEL453DWL","short_pith_number":"pith:FGRNTKQR","schema_version":"1.0","canonical_sha256":"29a2d9aa11368a751ad091c8be7763b2ea916e3c273a5221cda55c64b021979c","source":{"kind":"arxiv","id":"2602.03664","version":3},"attestation_state":"computed","paper":{"title":"Mitigating Conversational Inertia in Multi-Turn Agents","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.AI","authors_text":"Changhua Meng, Linchao Zhu, Shuheng Shen, Yang Wan, Zheng Cao, Zhengwen Zeng, Zhenhao Zhang","submitted_at":"2026-02-03T15:47:32Z","abstract_excerpt":"Large language models excel as few-shot learners when provided with appropriate demonstrations, yet this strength becomes problematic in multiturn agent scenarios, where LLMs erroneously mimic their own previous responses as few-shot examples. Through attention analysis, we identify conversational inertia, a phenomenon where models exhibit strong diagonal attention to previous responses, which is associated with imitation bias that constrains exploration. This reveals a tension when transforming few-shot LLMs into agents: longer context enriches environmental feedback for exploitation, yet als"},"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":"2602.03664","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-02-03T15:47:32Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"31bf3ccce1d1f2cd182bd9e408bafefb03368227591194ec225ee6f3ba90aabd","abstract_canon_sha256":"a3b39df357393f77afc904dc2b594adc57d15772b0aba41c869e5c52f4b9b654"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:05:41.545622Z","signature_b64":"f2Gy77WE29COShQ/LoKKypCBhyRHy8nUsstkRJi9u5Rz7DoZmvYMfFijDFsI1VCFJh2trdx/SDwp+CAeQRBGCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"29a2d9aa11368a751ad091c8be7763b2ea916e3c273a5221cda55c64b021979c","last_reissued_at":"2026-05-20T00:05:41.544928Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:05:41.544928Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Mitigating Conversational Inertia in Multi-Turn Agents","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.AI","authors_text":"Changhua Meng, Linchao Zhu, Shuheng Shen, Yang Wan, Zheng Cao, Zhengwen Zeng, Zhenhao Zhang","submitted_at":"2026-02-03T15:47:32Z","abstract_excerpt":"Large language models excel as few-shot learners when provided with appropriate demonstrations, yet this strength becomes problematic in multiturn agent scenarios, where LLMs erroneously mimic their own previous responses as few-shot examples. Through attention analysis, we identify conversational inertia, a phenomenon where models exhibit strong diagonal attention to previous responses, which is associated with imitation bias that constrains exploration. This reveals a tension when transforming few-shot LLMs into agents: longer context enriches environmental feedback for exploitation, yet als"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2602.03664","kind":"arxiv","version":3},"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/2602.03664/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":"2602.03664","created_at":"2026-05-20T00:05:41.545042+00:00"},{"alias_kind":"arxiv_version","alias_value":"2602.03664v3","created_at":"2026-05-20T00:05:41.545042+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2602.03664","created_at":"2026-05-20T00:05:41.545042+00:00"},{"alias_kind":"pith_short_12","alias_value":"FGRNTKQRG2FH","created_at":"2026-05-20T00:05:41.545042+00:00"},{"alias_kind":"pith_short_16","alias_value":"FGRNTKQRG2FHKGWQ","created_at":"2026-05-20T00:05:41.545042+00:00"},{"alias_kind":"pith_short_8","alias_value":"FGRNTKQR","created_at":"2026-05-20T00:05:41.545042+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2605.22714","citing_title":"AMEL: Accumulated Message Effects on LLM Judgments","ref_index":30,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/FGRNTKQRG2FHKGWQSHEL453DWL","json":"https://pith.science/pith/FGRNTKQRG2FHKGWQSHEL453DWL.json","graph_json":"https://pith.science/api/pith-number/FGRNTKQRG2FHKGWQSHEL453DWL/graph.json","events_json":"https://pith.science/api/pith-number/FGRNTKQRG2FHKGWQSHEL453DWL/events.json","paper":"https://pith.science/paper/FGRNTKQR"},"agent_actions":{"view_html":"https://pith.science/pith/FGRNTKQRG2FHKGWQSHEL453DWL","download_json":"https://pith.science/pith/FGRNTKQRG2FHKGWQSHEL453DWL.json","view_paper":"https://pith.science/paper/FGRNTKQR","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2602.03664&json=true","fetch_graph":"https://pith.science/api/pith-number/FGRNTKQRG2FHKGWQSHEL453DWL/graph.json","fetch_events":"https://pith.science/api/pith-number/FGRNTKQRG2FHKGWQSHEL453DWL/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/FGRNTKQRG2FHKGWQSHEL453DWL/action/timestamp_anchor","attest_storage":"https://pith.science/pith/FGRNTKQRG2FHKGWQSHEL453DWL/action/storage_attestation","attest_author":"https://pith.science/pith/FGRNTKQRG2FHKGWQSHEL453DWL/action/author_attestation","sign_citation":"https://pith.science/pith/FGRNTKQRG2FHKGWQSHEL453DWL/action/citation_signature","submit_replication":"https://pith.science/pith/FGRNTKQRG2FHKGWQSHEL453DWL/action/replication_record"}},"created_at":"2026-05-20T00:05:41.545042+00:00","updated_at":"2026-05-20T00:05:41.545042+00:00"}