{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:HWBWMP5YXRRWOQTVRXRUMCZTLK","short_pith_number":"pith:HWBWMP5Y","canonical_record":{"source":{"id":"2605.13369","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-13T11:27:40Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"129407ebd1cd3bcd2614d4223dc14cb773fdb0c6fff23a384f258abcd960d544","abstract_canon_sha256":"f8740c8dc59a74afea06d0eb91d97dd84ca9a0c9f9fcf3a24315c6ad7d783b4a"},"schema_version":"1.0"},"canonical_sha256":"3d83663fb8bc636742758de3460b335a93eb726f3a76521530b9bf4ec57bae6d","source":{"kind":"arxiv","id":"2605.13369","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.13369","created_at":"2026-05-18T02:44:47Z"},{"alias_kind":"arxiv_version","alias_value":"2605.13369v2","created_at":"2026-05-18T02:44:47Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.13369","created_at":"2026-05-18T02:44:47Z"},{"alias_kind":"pith_short_12","alias_value":"HWBWMP5YXRRW","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"HWBWMP5YXRRWOQTV","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"HWBWMP5Y","created_at":"2026-05-18T12:33:37Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:HWBWMP5YXRRWOQTVRXRUMCZTLK","target":"record","payload":{"canonical_record":{"source":{"id":"2605.13369","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-13T11:27:40Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"129407ebd1cd3bcd2614d4223dc14cb773fdb0c6fff23a384f258abcd960d544","abstract_canon_sha256":"f8740c8dc59a74afea06d0eb91d97dd84ca9a0c9f9fcf3a24315c6ad7d783b4a"},"schema_version":"1.0"},"canonical_sha256":"3d83663fb8bc636742758de3460b335a93eb726f3a76521530b9bf4ec57bae6d","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T02:44:47.996582Z","signature_b64":"Q6Wwdeq1A8pB46V12doBw8fn7NvaVrkN3TCJzffcrf4+4uXrarF6uejmFShh2aO5l3TxTIyXrHjSKYpMFmYkAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3d83663fb8bc636742758de3460b335a93eb726f3a76521530b9bf4ec57bae6d","last_reissued_at":"2026-05-18T02:44:47.996156Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T02:44:47.996156Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2605.13369","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-18T02:44:47Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"B6hGdCJ5EFjqUrHAs8qsUyrYreFb7UXNHW6i6RQJyqnAfGLXmcFz0F9dlnGnmkgeUk/KPmUNHG8+WtA9OEhMBQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-31T00:12:31.426494Z"},"content_sha256":"768de161dc436a5d28fdcbffe653fdec1a14bc60f58161b25801b0597d279301","schema_version":"1.0","event_id":"sha256:768de161dc436a5d28fdcbffe653fdec1a14bc60f58161b25801b0597d279301"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:HWBWMP5YXRRWOQTVRXRUMCZTLK","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Query-Conditioned Test-Time Self-Training for Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Large language models can adapt their own parameters during inference by generating training examples directly from the input query.","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Chaehee Song, Changick Kim, Doyi Kim, Minseok Seo, Yeeun Seong","submitted_at":"2026-05-13T11:27:40Z","abstract_excerpt":"Large language models (LLMs) are typically deployed with fixed parameters, and their performance is often improved by allocating more computation at inference time. While such test-time scaling can be effective, it cannot correct model misconceptions or adapt the model to the specific structure of an individual query. Test-time optimization addresses this limitation by enabling parameter updates during inference, but existing approaches either rely on external data or optimize generic self-supervised objectives that lack query-specific alignment. In this work, we propose Query-Conditioned Test"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"QueST generates such query-conditioned pairs and uses them as supervision for parameter-efficient fine-tuning at test time. The adapted model is then used to produce the final answer, enabling query-specific adaptation without any external data.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"the input query itself encodes latent signals sufficient for constructing structurally related problem--solution pairs","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"QueST adapts LLMs at test time by generating query-specific problem-solution pairs for self-supervised fine-tuning, improving reasoning performance without external data.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Large language models can adapt their own parameters during inference by generating training examples directly from the input query.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"b88bff5c21af8a3ba49945f1935cb529374619927cb482b0062bb67fce8e2df1"},"source":{"id":"2605.13369","kind":"arxiv","version":2},"verdict":{"id":"19d914b4-30f0-4f16-a522-95bf33f7836c","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-15T05:50:41.544249Z","strongest_claim":"QueST generates such query-conditioned pairs and uses them as supervision for parameter-efficient fine-tuning at test time. The adapted model is then used to produce the final answer, enabling query-specific adaptation without any external data.","one_line_summary":"QueST adapts LLMs at test time by generating query-specific problem-solution pairs for self-supervised fine-tuning, improving reasoning performance without external data.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"the input query itself encodes latent signals sufficient for constructing structurally related problem--solution pairs","pith_extraction_headline":"Large language models can adapt their own parameters during inference by generating training examples directly from the input query."},"references":{"count":42,"sample":[{"doi":"","year":2023,"title":"GPT-4 Technical Report","work_id":"b928e041-6991-4c08-8c81-0359e4097c7b","ref_index":1,"cited_arxiv_id":"2303.08774","is_internal_anchor":true},{"doi":"","year":2021,"title":"Training Verifiers to Solve Math Word Problems","work_id":"acab1aa8-b4d6-40e0-a3ee-25341701dca2","ref_index":2,"cited_arxiv_id":"2110.14168","is_internal_anchor":true},{"doi":"","year":2026,"title":"In-Place Test-Time Training","work_id":"7edff7e1-6263-47c6-afd5-e77029c3948a","ref_index":3,"cited_arxiv_id":"2604.06169","is_internal_anchor":true},{"doi":"","year":2025,"title":"DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning","work_id":"e6b75ad5-2877-4168-97c8-710407094d20","ref_index":4,"cited_arxiv_id":"2501.12948","is_internal_anchor":true},{"doi":"","year":2023,"title":"Test-time training on nearest neighbors for large language models.arXiv preprint arXiv:2305.18466","work_id":"88670479-3b0a-4a44-8a0d-d11feb3aefdd","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":42,"snapshot_sha256":"7c533a3b618ccf95b732bc4a5aae4b75fba986ea15a8346514bd0429d8cddd2b","internal_anchors":14},"formal_canon":{"evidence_count":2,"snapshot_sha256":"f3b93275f7df4723275dca8b4d28e92e0e092133e3be7eff10726c1eb13d2568"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"verdict_id":"19d914b4-30f0-4f16-a522-95bf33f7836c"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-18T02:44:47Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"v4CVqoTNMvQwbB8qLTfArySt8iksd+hwo236jIBo0InyiLZx6P4mxoNzEpUyv4dZSa7KTpTyc4fbB3h/Ur5pCw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-31T00:12:31.427029Z"},"content_sha256":"a2cf3e09a5f1d9c55a147ff9e9b5f18ef498d80e58f29ba3e454f604e5872eeb","schema_version":"1.0","event_id":"sha256:a2cf3e09a5f1d9c55a147ff9e9b5f18ef498d80e58f29ba3e454f604e5872eeb"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/HWBWMP5YXRRWOQTVRXRUMCZTLK/bundle.json","state_url":"https://pith.science/pith/HWBWMP5YXRRWOQTVRXRUMCZTLK/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/HWBWMP5YXRRWOQTVRXRUMCZTLK/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-05-31T00:12:31Z","links":{"resolver":"https://pith.science/pith/HWBWMP5YXRRWOQTVRXRUMCZTLK","bundle":"https://pith.science/pith/HWBWMP5YXRRWOQTVRXRUMCZTLK/bundle.json","state":"https://pith.science/pith/HWBWMP5YXRRWOQTVRXRUMCZTLK/state.json","well_known_bundle":"https://pith.science/.well-known/pith/HWBWMP5YXRRWOQTVRXRUMCZTLK/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:HWBWMP5YXRRWOQTVRXRUMCZTLK","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"f8740c8dc59a74afea06d0eb91d97dd84ca9a0c9f9fcf3a24315c6ad7d783b4a","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-13T11:27:40Z","title_canon_sha256":"129407ebd1cd3bcd2614d4223dc14cb773fdb0c6fff23a384f258abcd960d544"},"schema_version":"1.0","source":{"id":"2605.13369","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.13369","created_at":"2026-05-18T02:44:47Z"},{"alias_kind":"arxiv_version","alias_value":"2605.13369v2","created_at":"2026-05-18T02:44:47Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.13369","created_at":"2026-05-18T02:44:47Z"},{"alias_kind":"pith_short_12","alias_value":"HWBWMP5YXRRW","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"HWBWMP5YXRRWOQTV","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"HWBWMP5Y","created_at":"2026-05-18T12:33:37Z"}],"graph_snapshots":[{"event_id":"sha256:a2cf3e09a5f1d9c55a147ff9e9b5f18ef498d80e58f29ba3e454f604e5872eeb","target":"graph","created_at":"2026-05-18T02:44:47Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"QueST generates such query-conditioned pairs and uses them as supervision for parameter-efficient fine-tuning at test time. The adapted model is then used to produce the final answer, enabling query-specific adaptation without any external data."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"the input query itself encodes latent signals sufficient for constructing structurally related problem--solution pairs"},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"QueST adapts LLMs at test time by generating query-specific problem-solution pairs for self-supervised fine-tuning, improving reasoning performance without external data."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Large language models can adapt their own parameters during inference by generating training examples directly from the input query."}],"snapshot_sha256":"b88bff5c21af8a3ba49945f1935cb529374619927cb482b0062bb67fce8e2df1"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"f3b93275f7df4723275dca8b4d28e92e0e092133e3be7eff10726c1eb13d2568"},"paper":{"abstract_excerpt":"Large language models (LLMs) are typically deployed with fixed parameters, and their performance is often improved by allocating more computation at inference time. While such test-time scaling can be effective, it cannot correct model misconceptions or adapt the model to the specific structure of an individual query. Test-time optimization addresses this limitation by enabling parameter updates during inference, but existing approaches either rely on external data or optimize generic self-supervised objectives that lack query-specific alignment. In this work, we propose Query-Conditioned Test","authors_text":"Chaehee Song, Changick Kim, Doyi Kim, Minseok Seo, Yeeun Seong","cross_cats":["cs.AI","cs.LG"],"headline":"Large language models can adapt their own parameters during inference by generating training examples directly from the input query.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-13T11:27:40Z","title":"Query-Conditioned Test-Time Self-Training for Large Language Models"},"references":{"count":42,"internal_anchors":14,"resolved_work":42,"sample":[{"cited_arxiv_id":"2303.08774","doi":"","is_internal_anchor":true,"ref_index":1,"title":"GPT-4 Technical Report","work_id":"b928e041-6991-4c08-8c81-0359e4097c7b","year":2023},{"cited_arxiv_id":"2110.14168","doi":"","is_internal_anchor":true,"ref_index":2,"title":"Training Verifiers to Solve Math Word Problems","work_id":"acab1aa8-b4d6-40e0-a3ee-25341701dca2","year":2021},{"cited_arxiv_id":"2604.06169","doi":"","is_internal_anchor":true,"ref_index":3,"title":"In-Place Test-Time Training","work_id":"7edff7e1-6263-47c6-afd5-e77029c3948a","year":2026},{"cited_arxiv_id":"2501.12948","doi":"","is_internal_anchor":true,"ref_index":4,"title":"DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning","work_id":"e6b75ad5-2877-4168-97c8-710407094d20","year":2025},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"Test-time training on nearest neighbors for large language models.arXiv preprint arXiv:2305.18466","work_id":"88670479-3b0a-4a44-8a0d-d11feb3aefdd","year":2023}],"snapshot_sha256":"7c533a3b618ccf95b732bc4a5aae4b75fba986ea15a8346514bd0429d8cddd2b"},"source":{"id":"2605.13369","kind":"arxiv","version":2},"verdict":{"created_at":"2026-05-15T05:50:41.544249Z","id":"19d914b4-30f0-4f16-a522-95bf33f7836c","model_set":{"reader":"grok-4.3"},"one_line_summary":"QueST adapts LLMs at test time by generating query-specific problem-solution pairs for self-supervised fine-tuning, improving reasoning performance without external data.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Large language models can adapt their own parameters during inference by generating training examples directly from the input query.","strongest_claim":"QueST generates such query-conditioned pairs and uses them as supervision for parameter-efficient fine-tuning at test time. The adapted model is then used to produce the final answer, enabling query-specific adaptation without any external data.","weakest_assumption":"the input query itself encodes latent signals sufficient for constructing structurally related problem--solution pairs"}},"verdict_id":"19d914b4-30f0-4f16-a522-95bf33f7836c"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:768de161dc436a5d28fdcbffe653fdec1a14bc60f58161b25801b0597d279301","target":"record","created_at":"2026-05-18T02:44:47Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"f8740c8dc59a74afea06d0eb91d97dd84ca9a0c9f9fcf3a24315c6ad7d783b4a","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-13T11:27:40Z","title_canon_sha256":"129407ebd1cd3bcd2614d4223dc14cb773fdb0c6fff23a384f258abcd960d544"},"schema_version":"1.0","source":{"id":"2605.13369","kind":"arxiv","version":2}},"canonical_sha256":"3d83663fb8bc636742758de3460b335a93eb726f3a76521530b9bf4ec57bae6d","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"3d83663fb8bc636742758de3460b335a93eb726f3a76521530b9bf4ec57bae6d","first_computed_at":"2026-05-18T02:44:47.996156Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-18T02:44:47.996156Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"Q6Wwdeq1A8pB46V12doBw8fn7NvaVrkN3TCJzffcrf4+4uXrarF6uejmFShh2aO5l3TxTIyXrHjSKYpMFmYkAg==","signature_status":"signed_v1","signed_at":"2026-05-18T02:44:47.996582Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.13369","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:768de161dc436a5d28fdcbffe653fdec1a14bc60f58161b25801b0597d279301","sha256:a2cf3e09a5f1d9c55a147ff9e9b5f18ef498d80e58f29ba3e454f604e5872eeb"],"state_sha256":"4632581b5d882cb743110c3b66d289b44ebdce69b75f99632ce52903f8b8701f"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Lh1gfz6bJE7QveDd6u7767Qz9gee/J4RXBfjqvBhdLXVCqWK9/gizfKnt7ooxVzcbYuC7Djjx4IUhDRkxroGDw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-31T00:12:31.429918Z","bundle_sha256":"45403e9a48eccb20debba2bc221da72ea82069d2a42a84df835dc10593cf5b5c"}}