{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:TWZYCI6YROJW5DCR5JO4BVZL3V","short_pith_number":"pith:TWZYCI6Y","schema_version":"1.0","canonical_sha256":"9db38123d88b936e8c51ea5dc0d72bdd6d36348fa2e3f18ba3b26047ac6471d2","source":{"kind":"arxiv","id":"2606.09662","version":1},"attestation_state":"computed","paper":{"title":"When Built-in Thinking Helps and Hurts: Constraint-Level Error Shifts in Instruction Following","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Sai Adith Senthil Kumar","submitted_at":"2026-06-08T15:45:04Z","abstract_excerpt":"Large reasoning models (LRMs) often improve math and coding performance, but their effect on instruction following is unclear. We study IFEval with Qwen3 models (1.7B-32B), using same-weights Thinking ON/OFF controls; four Hunyuan models provide directional cross-family support. Aggregate pass-rate changes are small (-0.55 to -3.52 pp), yet 10-20% of prompts switch between pass and fail across modes, suggesting that thinking changes the pattern of errors--some prompts improve while others worsen--rather than uniformly degrading performance. Under a post-hoc Qwen3-derived grouping, constraint t"},"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":"2606.09662","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-06-08T15:45:04Z","cross_cats_sorted":[],"title_canon_sha256":"c5e88714ffb7f33a824849147f15a86255619af6ae9cd38b3e2aa38587269089","abstract_canon_sha256":"e310d13b63848185f745291760370a6f1ee410b37c98d3f9bfee69ee5b0a8e0a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-09T02:09:02.218202Z","signature_b64":"QQhysdzEqNKC1GOc1enbcVKcQca5p2iAyiuvK+9ITnSyc04ND7fD35y4bDgknOCcuPdPpP8KvUC/bKk7pZGCCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9db38123d88b936e8c51ea5dc0d72bdd6d36348fa2e3f18ba3b26047ac6471d2","last_reissued_at":"2026-06-09T02:09:02.217374Z","signature_status":"signed_v1","first_computed_at":"2026-06-09T02:09:02.217374Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"When Built-in Thinking Helps and Hurts: Constraint-Level Error Shifts in Instruction Following","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Sai Adith Senthil Kumar","submitted_at":"2026-06-08T15:45:04Z","abstract_excerpt":"Large reasoning models (LRMs) often improve math and coding performance, but their effect on instruction following is unclear. We study IFEval with Qwen3 models (1.7B-32B), using same-weights Thinking ON/OFF controls; four Hunyuan models provide directional cross-family support. Aggregate pass-rate changes are small (-0.55 to -3.52 pp), yet 10-20% of prompts switch between pass and fail across modes, suggesting that thinking changes the pattern of errors--some prompts improve while others worsen--rather than uniformly degrading performance. Under a post-hoc Qwen3-derived grouping, constraint t"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.09662","kind":"arxiv","version":1},"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/2606.09662/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":"2606.09662","created_at":"2026-06-09T02:09:02.217515+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.09662v1","created_at":"2026-06-09T02:09:02.217515+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.09662","created_at":"2026-06-09T02:09:02.217515+00:00"},{"alias_kind":"pith_short_12","alias_value":"TWZYCI6YROJW","created_at":"2026-06-09T02:09:02.217515+00:00"},{"alias_kind":"pith_short_16","alias_value":"TWZYCI6YROJW5DCR","created_at":"2026-06-09T02:09:02.217515+00:00"},{"alias_kind":"pith_short_8","alias_value":"TWZYCI6Y","created_at":"2026-06-09T02:09:02.217515+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/TWZYCI6YROJW5DCR5JO4BVZL3V","json":"https://pith.science/pith/TWZYCI6YROJW5DCR5JO4BVZL3V.json","graph_json":"https://pith.science/api/pith-number/TWZYCI6YROJW5DCR5JO4BVZL3V/graph.json","events_json":"https://pith.science/api/pith-number/TWZYCI6YROJW5DCR5JO4BVZL3V/events.json","paper":"https://pith.science/paper/TWZYCI6Y"},"agent_actions":{"view_html":"https://pith.science/pith/TWZYCI6YROJW5DCR5JO4BVZL3V","download_json":"https://pith.science/pith/TWZYCI6YROJW5DCR5JO4BVZL3V.json","view_paper":"https://pith.science/paper/TWZYCI6Y","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.09662&json=true","fetch_graph":"https://pith.science/api/pith-number/TWZYCI6YROJW5DCR5JO4BVZL3V/graph.json","fetch_events":"https://pith.science/api/pith-number/TWZYCI6YROJW5DCR5JO4BVZL3V/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/TWZYCI6YROJW5DCR5JO4BVZL3V/action/timestamp_anchor","attest_storage":"https://pith.science/pith/TWZYCI6YROJW5DCR5JO4BVZL3V/action/storage_attestation","attest_author":"https://pith.science/pith/TWZYCI6YROJW5DCR5JO4BVZL3V/action/author_attestation","sign_citation":"https://pith.science/pith/TWZYCI6YROJW5DCR5JO4BVZL3V/action/citation_signature","submit_replication":"https://pith.science/pith/TWZYCI6YROJW5DCR5JO4BVZL3V/action/replication_record"}},"created_at":"2026-06-09T02:09:02.217515+00:00","updated_at":"2026-06-09T02:09:02.217515+00:00"}