{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:XKNODIBJ2LJGWKS2SXX7KWJON2","short_pith_number":"pith:XKNODIBJ","canonical_record":{"source":{"id":"2607.23153","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-07-25T11:13:52Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"9ba589bec8ff077c48f70dff96808e237179d0d48f04fe2fb83a03ac7660e004","abstract_canon_sha256":"d91142588680578a815bf590dc4c19690a6d4ca55e4bd41457a9abcde14f1620"},"schema_version":"1.0"},"canonical_sha256":"ba9ae1a029d2d26b2a5a95eff5592e6eaa900f0b39aa1670de70803f98d393dc","source":{"kind":"arxiv","id":"2607.23153","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2607.23153","created_at":"2026-07-28T01:22:09Z"},{"alias_kind":"arxiv_version","alias_value":"2607.23153v1","created_at":"2026-07-28T01:22:09Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.23153","created_at":"2026-07-28T01:22:09Z"},{"alias_kind":"pith_short_12","alias_value":"XKNODIBJ2LJG","created_at":"2026-07-28T01:22:09Z"},{"alias_kind":"pith_short_16","alias_value":"XKNODIBJ2LJGWKS2","created_at":"2026-07-28T01:22:09Z"},{"alias_kind":"pith_short_8","alias_value":"XKNODIBJ","created_at":"2026-07-28T01:22:09Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:XKNODIBJ2LJGWKS2SXX7KWJON2","target":"record","payload":{"canonical_record":{"source":{"id":"2607.23153","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-07-25T11:13:52Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"9ba589bec8ff077c48f70dff96808e237179d0d48f04fe2fb83a03ac7660e004","abstract_canon_sha256":"d91142588680578a815bf590dc4c19690a6d4ca55e4bd41457a9abcde14f1620"},"schema_version":"1.0"},"canonical_sha256":"ba9ae1a029d2d26b2a5a95eff5592e6eaa900f0b39aa1670de70803f98d393dc","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-28T01:22:09.662687Z","signature_b64":"bQHCxFSTi6kWd9LDEKKsqzmw4Ay3XTCIZBpHUgW0dNwaD5Rigw2I6fxBZJGGBMU/3N2FF5skhw56cHOWvjK/DQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ba9ae1a029d2d26b2a5a95eff5592e6eaa900f0b39aa1670de70803f98d393dc","last_reissued_at":"2026-07-28T01:22:09.661809Z","signature_status":"signed_v1","first_computed_at":"2026-07-28T01:22:09.661809Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2607.23153","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-28T01:22:09Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Sxp0uc5T7z9TPB/0vyM2wd2nYq4zTjR+IjwjkZaW2sh89z0we05EZjGtDCLO4NC4z14r+Hh8OI1CEfHPMdZsDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-10T16:31:11.039537Z"},"content_sha256":"8b219ea59263e16b523b616ff25e947d96c6917b504cce534749dd66a0a64811","schema_version":"1.0","event_id":"sha256:8b219ea59263e16b523b616ff25e947d96c6917b504cce534749dd66a0a64811"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:XKNODIBJ2LJGWKS2SXX7KWJON2","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"In-Context Learning as Implicit Policy Gradient","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Masahiro Kaneko, Timothy Baldwin","submitted_at":"2026-07-25T11:13:52Z","abstract_excerpt":"Recent work has shown that large language models (LLMs) can iteratively improve their outputs by incorporating generated samples and their corresponding evaluation scores as in-context examples. Despite these empirical findings, the theoretical foundations underlying this phenomenon remain poorly understood. In this paper, we show that score-conditioned In-Context Learning (ICL) admits a structural correspondence to policy gradient optimization. We first provide a constructive proof that self-attention mechanisms can implement reward-weighted aggregation analogous to the REINFORCE algorithm un"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.23153","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/2607.23153/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-28T01:22:09Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"0f6vX+roC6OKG5TInE1sTTBlX+JEMKnq5yBGCBZ8ha+NEUeOr/26vB/dfkPHh8OvxoA/fEMIq9Wrdj0UxpfeAg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-10T16:31:11.040053Z"},"content_sha256":"0066015230201971c199fc1f287d4192e2b4070acfb1e05ec0a9a2437260b33d","schema_version":"1.0","event_id":"sha256:0066015230201971c199fc1f287d4192e2b4070acfb1e05ec0a9a2437260b33d"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/XKNODIBJ2LJGWKS2SXX7KWJON2/bundle.json","state_url":"https://pith.science/pith/XKNODIBJ2LJGWKS2SXX7KWJON2/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/XKNODIBJ2LJGWKS2SXX7KWJON2/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":"2026-08-10T16:31:11Z","links":{"resolver":"https://pith.science/pith/XKNODIBJ2LJGWKS2SXX7KWJON2","bundle":"https://pith.science/pith/XKNODIBJ2LJGWKS2SXX7KWJON2/bundle.json","state":"https://pith.science/pith/XKNODIBJ2LJGWKS2SXX7KWJON2/state.json","well_known_bundle":"https://pith.science/.well-known/pith/XKNODIBJ2LJGWKS2SXX7KWJON2/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:XKNODIBJ2LJGWKS2SXX7KWJON2","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":"d91142588680578a815bf590dc4c19690a6d4ca55e4bd41457a9abcde14f1620","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-07-25T11:13:52Z","title_canon_sha256":"9ba589bec8ff077c48f70dff96808e237179d0d48f04fe2fb83a03ac7660e004"},"schema_version":"1.0","source":{"id":"2607.23153","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2607.23153","created_at":"2026-07-28T01:22:09Z"},{"alias_kind":"arxiv_version","alias_value":"2607.23153v1","created_at":"2026-07-28T01:22:09Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.23153","created_at":"2026-07-28T01:22:09Z"},{"alias_kind":"pith_short_12","alias_value":"XKNODIBJ2LJG","created_at":"2026-07-28T01:22:09Z"},{"alias_kind":"pith_short_16","alias_value":"XKNODIBJ2LJGWKS2","created_at":"2026-07-28T01:22:09Z"},{"alias_kind":"pith_short_8","alias_value":"XKNODIBJ","created_at":"2026-07-28T01:22:09Z"}],"graph_snapshots":[{"event_id":"sha256:0066015230201971c199fc1f287d4192e2b4070acfb1e05ec0a9a2437260b33d","target":"graph","created_at":"2026-07-28T01:22:09Z","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":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2607.23153/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Recent work has shown that large language models (LLMs) can iteratively improve their outputs by incorporating generated samples and their corresponding evaluation scores as in-context examples. Despite these empirical findings, the theoretical foundations underlying this phenomenon remain poorly understood. In this paper, we show that score-conditioned In-Context Learning (ICL) admits a structural correspondence to policy gradient optimization. We first provide a constructive proof that self-attention mechanisms can implement reward-weighted aggregation analogous to the REINFORCE algorithm un","authors_text":"Masahiro Kaneko, Timothy Baldwin","cross_cats":["cs.AI","cs.CL"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-07-25T11:13:52Z","title":"In-Context Learning as Implicit Policy Gradient"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.23153","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:8b219ea59263e16b523b616ff25e947d96c6917b504cce534749dd66a0a64811","target":"record","created_at":"2026-07-28T01:22:09Z","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":"d91142588680578a815bf590dc4c19690a6d4ca55e4bd41457a9abcde14f1620","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-07-25T11:13:52Z","title_canon_sha256":"9ba589bec8ff077c48f70dff96808e237179d0d48f04fe2fb83a03ac7660e004"},"schema_version":"1.0","source":{"id":"2607.23153","kind":"arxiv","version":1}},"canonical_sha256":"ba9ae1a029d2d26b2a5a95eff5592e6eaa900f0b39aa1670de70803f98d393dc","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"ba9ae1a029d2d26b2a5a95eff5592e6eaa900f0b39aa1670de70803f98d393dc","first_computed_at":"2026-07-28T01:22:09.661809Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-28T01:22:09.661809Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"bQHCxFSTi6kWd9LDEKKsqzmw4Ay3XTCIZBpHUgW0dNwaD5Rigw2I6fxBZJGGBMU/3N2FF5skhw56cHOWvjK/DQ==","signature_status":"signed_v1","signed_at":"2026-07-28T01:22:09.662687Z","signed_message":"canonical_sha256_bytes"},"source_id":"2607.23153","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:8b219ea59263e16b523b616ff25e947d96c6917b504cce534749dd66a0a64811","sha256:0066015230201971c199fc1f287d4192e2b4070acfb1e05ec0a9a2437260b33d"],"state_sha256":"b7307ebcde81a02e0333a7d029032d78a153d520d34e16d0156ffa4a6d480bc6"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"7vXVjWQmiCDpr1y9onSfhwx68CZx5f8PshqA50X9688BrX9HAoFXwQh4/62qskZ9FMcOLfFQZsGWSU/bDRH+AQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-10T16:31:11.046265Z","bundle_sha256":"e4782a096855bef3dac0888feab9a3428ea109d0b249bac573c417ca51a4c428"}}