{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:KYOPUF3IL6FAAFDTGMA5JDZ7AN","short_pith_number":"pith:KYOPUF3I","canonical_record":{"source":{"id":"2605.25459","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-25T06:09:04Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"accf50ee254c2b413f710d56039497abfad63fc4722194d5d3cae8f57407a909","abstract_canon_sha256":"cfd9dc988b4bd1813b91a12909df4994115524afac298e22ff9b7cbfb04c0a54"},"schema_version":"1.0"},"canonical_sha256":"561cfa17685f8a0014733301d48f3f0345abacffd34a54ce8871e9c7d8924459","source":{"kind":"arxiv","id":"2605.25459","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.25459","created_at":"2026-05-26T02:04:36Z"},{"alias_kind":"arxiv_version","alias_value":"2605.25459v1","created_at":"2026-05-26T02:04:36Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.25459","created_at":"2026-05-26T02:04:36Z"},{"alias_kind":"pith_short_12","alias_value":"KYOPUF3IL6FA","created_at":"2026-05-26T02:04:36Z"},{"alias_kind":"pith_short_16","alias_value":"KYOPUF3IL6FAAFDT","created_at":"2026-05-26T02:04:36Z"},{"alias_kind":"pith_short_8","alias_value":"KYOPUF3I","created_at":"2026-05-26T02:04:36Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:KYOPUF3IL6FAAFDTGMA5JDZ7AN","target":"record","payload":{"canonical_record":{"source":{"id":"2605.25459","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-25T06:09:04Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"accf50ee254c2b413f710d56039497abfad63fc4722194d5d3cae8f57407a909","abstract_canon_sha256":"cfd9dc988b4bd1813b91a12909df4994115524afac298e22ff9b7cbfb04c0a54"},"schema_version":"1.0"},"canonical_sha256":"561cfa17685f8a0014733301d48f3f0345abacffd34a54ce8871e9c7d8924459","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-26T02:04:36.881072Z","signature_b64":"3BTqcdi2cJg4pebHBFjEtoy2if6cGJVGAdQBYPESUjfuHW0lV2W/ZVhnooRlapbO0+52nCn+QoWrfb3A5iUWBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"561cfa17685f8a0014733301d48f3f0345abacffd34a54ce8871e9c7d8924459","last_reissued_at":"2026-05-26T02:04:36.880204Z","signature_status":"signed_v1","first_computed_at":"2026-05-26T02:04:36.880204Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2605.25459","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":"2026-05-26T02:04:36Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"F3Zo4HiD9BCtdF+6l+sDFVz3TJYP/SKCkZQ1Nstrdqc8y6myXDCxaEPmVXifmlRzrapgLLqdjcV/8dfiIi3kAg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-29T18:17:41.789035Z"},"content_sha256":"9de0332401a09a102bcb35e008975da4b3046d634de9e571b414521e1e1866b8","schema_version":"1.0","event_id":"sha256:9de0332401a09a102bcb35e008975da4b3046d634de9e571b414521e1e1866b8"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:KYOPUF3IL6FAAFDTGMA5JDZ7AN","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"From Simulation to Enaction: Post-trained language models recognize and react to their own generations","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Asvin G., Jack Lindsey","submitted_at":"2026-05-25T06:09:04Z","abstract_excerpt":"Language models are pretrained as passive predictors with no incentive to model the consequences of their own outputs. Post-training changes this: a model producing its own responses can benefit from recognizing that it is on-policy. We present evidence that post-trained models recognize their on-policy generations, and this recognition is implicitly encoded in their output distributions. In particular, on-policy output distribution entropy is 3--4$\\times$ lower than off-policy entropy, across model families and size classes. We trace part of this effect to an internal representation of input "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.25459","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/2605.25459/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":"2026-05-26T02:04:36Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"+y7dSzbuOPgDson3JL5WRd6BOWWhw3zwbOKTAqbAjD1Diq9yECc1ZnfqyT9/EhSJ5f5h3nm4hu4zSnUEOuIFBQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-29T18:17:41.789420Z"},"content_sha256":"563d6cfb21515b64965d1508e3ee6c397572d04d3db26067397cb2c1a9a6d59c","schema_version":"1.0","event_id":"sha256:563d6cfb21515b64965d1508e3ee6c397572d04d3db26067397cb2c1a9a6d59c"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/KYOPUF3IL6FAAFDTGMA5JDZ7AN/bundle.json","state_url":"https://pith.science/pith/KYOPUF3IL6FAAFDTGMA5JDZ7AN/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/KYOPUF3IL6FAAFDTGMA5JDZ7AN/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":"2026-06-29T18:17:41Z","links":{"resolver":"https://pith.science/pith/KYOPUF3IL6FAAFDTGMA5JDZ7AN","bundle":"https://pith.science/pith/KYOPUF3IL6FAAFDTGMA5JDZ7AN/bundle.json","state":"https://pith.science/pith/KYOPUF3IL6FAAFDTGMA5JDZ7AN/state.json","well_known_bundle":"https://pith.science/.well-known/pith/KYOPUF3IL6FAAFDTGMA5JDZ7AN/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:KYOPUF3IL6FAAFDTGMA5JDZ7AN","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":"cfd9dc988b4bd1813b91a12909df4994115524afac298e22ff9b7cbfb04c0a54","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-25T06:09:04Z","title_canon_sha256":"accf50ee254c2b413f710d56039497abfad63fc4722194d5d3cae8f57407a909"},"schema_version":"1.0","source":{"id":"2605.25459","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.25459","created_at":"2026-05-26T02:04:36Z"},{"alias_kind":"arxiv_version","alias_value":"2605.25459v1","created_at":"2026-05-26T02:04:36Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.25459","created_at":"2026-05-26T02:04:36Z"},{"alias_kind":"pith_short_12","alias_value":"KYOPUF3IL6FA","created_at":"2026-05-26T02:04:36Z"},{"alias_kind":"pith_short_16","alias_value":"KYOPUF3IL6FAAFDT","created_at":"2026-05-26T02:04:36Z"},{"alias_kind":"pith_short_8","alias_value":"KYOPUF3I","created_at":"2026-05-26T02:04:36Z"}],"graph_snapshots":[{"event_id":"sha256:563d6cfb21515b64965d1508e3ee6c397572d04d3db26067397cb2c1a9a6d59c","target":"graph","created_at":"2026-05-26T02:04:36Z","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/2605.25459/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Language models are pretrained as passive predictors with no incentive to model the consequences of their own outputs. Post-training changes this: a model producing its own responses can benefit from recognizing that it is on-policy. We present evidence that post-trained models recognize their on-policy generations, and this recognition is implicitly encoded in their output distributions. In particular, on-policy output distribution entropy is 3--4$\\times$ lower than off-policy entropy, across model families and size classes. We trace part of this effect to an internal representation of input ","authors_text":"Asvin G., Jack Lindsey","cross_cats":["cs.AI"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-25T06:09:04Z","title":"From Simulation to Enaction: Post-trained language models recognize and react to their own generations"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.25459","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":"sha256:9de0332401a09a102bcb35e008975da4b3046d634de9e571b414521e1e1866b8","target":"record","created_at":"2026-05-26T02:04:36Z","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":"cfd9dc988b4bd1813b91a12909df4994115524afac298e22ff9b7cbfb04c0a54","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-25T06:09:04Z","title_canon_sha256":"accf50ee254c2b413f710d56039497abfad63fc4722194d5d3cae8f57407a909"},"schema_version":"1.0","source":{"id":"2605.25459","kind":"arxiv","version":1}},"canonical_sha256":"561cfa17685f8a0014733301d48f3f0345abacffd34a54ce8871e9c7d8924459","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"561cfa17685f8a0014733301d48f3f0345abacffd34a54ce8871e9c7d8924459","first_computed_at":"2026-05-26T02:04:36.880204Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-26T02:04:36.880204Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"3BTqcdi2cJg4pebHBFjEtoy2if6cGJVGAdQBYPESUjfuHW0lV2W/ZVhnooRlapbO0+52nCn+QoWrfb3A5iUWBg==","signature_status":"signed_v1","signed_at":"2026-05-26T02:04:36.881072Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.25459","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:9de0332401a09a102bcb35e008975da4b3046d634de9e571b414521e1e1866b8","sha256:563d6cfb21515b64965d1508e3ee6c397572d04d3db26067397cb2c1a9a6d59c"],"state_sha256":"23c0e801815eba91689f4bde041b411b31e3a4fab715fd9ac8c0bd008c175bd0"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"VhAJcq2VadK0VXwAG8r6guYAgNZuElbmA0/g1FNjAyMiZ3DX1+7Pa5m7o6/Tg9dGlDRcUYutE6m9MSH4hKl6Dw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-29T18:17:41.791410Z","bundle_sha256":"94d74ceb36da4f3c17bc2a44f829756425303d02fcfb74216b01b2d440116177"}}