{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:AKUOQVQN52P2DJ67E2FONQFX7Q","short_pith_number":"pith:AKUOQVQN","canonical_record":{"source":{"id":"2501.19358","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-01-31T18:10:53Z","cross_cats_sorted":[],"title_canon_sha256":"5d44f02811377d4cb6672388eeeb2f70ccf2255126fa1f5b1d4a7aede10fd668","abstract_canon_sha256":"76aff074b5659055ae86044c4066f2650493af99142ff8971d6ed9d78d439499"},"schema_version":"1.0"},"canonical_sha256":"02a8e8560dee9fa1a7df268ae6c0b7fc0c650d2721f7bf85dec6e2d0747d48d0","source":{"kind":"arxiv","id":"2501.19358","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2501.19358","created_at":"2026-07-05T11:14:03Z"},{"alias_kind":"arxiv_version","alias_value":"2501.19358v3","created_at":"2026-07-05T11:14:03Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.19358","created_at":"2026-07-05T11:14:03Z"},{"alias_kind":"pith_short_12","alias_value":"AKUOQVQN52P2","created_at":"2026-07-05T11:14:03Z"},{"alias_kind":"pith_short_16","alias_value":"AKUOQVQN52P2DJ67","created_at":"2026-07-05T11:14:03Z"},{"alias_kind":"pith_short_8","alias_value":"AKUOQVQN","created_at":"2026-07-05T11:14:03Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:AKUOQVQN52P2DJ67E2FONQFX7Q","target":"record","payload":{"canonical_record":{"source":{"id":"2501.19358","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-01-31T18:10:53Z","cross_cats_sorted":[],"title_canon_sha256":"5d44f02811377d4cb6672388eeeb2f70ccf2255126fa1f5b1d4a7aede10fd668","abstract_canon_sha256":"76aff074b5659055ae86044c4066f2650493af99142ff8971d6ed9d78d439499"},"schema_version":"1.0"},"canonical_sha256":"02a8e8560dee9fa1a7df268ae6c0b7fc0c650d2721f7bf85dec6e2d0747d48d0","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:14:03.018776Z","signature_b64":"JMWcqfnbkwmbsv3n6H6BfcdY4qTU+T4J1sZdc/NM+jCs/10H1xO2oz7HPRyMYmzR25P1Y+W78tSHF20o2ZfFCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"02a8e8560dee9fa1a7df268ae6c0b7fc0c650d2721f7bf85dec6e2d0747d48d0","last_reissued_at":"2026-07-05T11:14:03.018222Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:14:03.018222Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2501.19358","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:14:03Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"qHsh0dcC8ERUAb2YBAc1Dp3YSyWeObVwIxl7KuM9/Zz1VMMyrB72uUzAz3I2KLpv4qFHBjVRDqv2aOc6moP4DA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-07T03:59:39.291656Z"},"content_sha256":"03029d4abd6d83321b86c882e347ae922a25c6ad5f208ea3f16b2832c42d3d93","schema_version":"1.0","event_id":"sha256:03029d4abd6d83321b86c882e347ae922a25c6ad5f208ea3f16b2832c42d3d93"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:AKUOQVQN52P2DJ67E2FONQFX7Q","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"The Energy Loss Phenomenon in RLHF: A New Perspective on Mitigating Reward Hacking","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Dacheng Tao, Lefei Zhang, Liang Ding, Sen Zhang, Yuchun Miao, Yuqi Zhang","submitted_at":"2025-01-31T18:10:53Z","abstract_excerpt":"This work identifies the Energy Loss Phenomenon in Reinforcement Learning from Human Feedback (RLHF) and its connection to reward hacking. Specifically, energy loss in the final layer of a Large Language Model (LLM) gradually increases during the RL process, with an excessive increase in energy loss characterizing reward hacking. Beyond empirical analysis, we further provide a theoretical foundation by proving that, under mild conditions, the increased energy loss reduces the upper bound of contextual relevance in LLMs, which is a critical aspect of reward hacking as the reduced contextual rel"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.19358","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/2501.19358/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":"2026-07-05T11:14:03Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"TRGb5RqY6X8f1zIilbX46vL53ioQCXEi7koeTqhVsHQIYWDMHmzHe3kDIJzqcrWPHtMosTFvOnKYBG3EAPpiCg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-07T03:59:39.292226Z"},"content_sha256":"27efa2cd7141183de9b2fc96dc2a97f287ea4695b6efef9c012aa991632f178b","schema_version":"1.0","event_id":"sha256:27efa2cd7141183de9b2fc96dc2a97f287ea4695b6efef9c012aa991632f178b"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/AKUOQVQN52P2DJ67E2FONQFX7Q/bundle.json","state_url":"https://pith.science/pith/AKUOQVQN52P2DJ67E2FONQFX7Q/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/AKUOQVQN52P2DJ67E2FONQFX7Q/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":"2026-08-07T03:59:39Z","links":{"resolver":"https://pith.science/pith/AKUOQVQN52P2DJ67E2FONQFX7Q","bundle":"https://pith.science/pith/AKUOQVQN52P2DJ67E2FONQFX7Q/bundle.json","state":"https://pith.science/pith/AKUOQVQN52P2DJ67E2FONQFX7Q/state.json","well_known_bundle":"https://pith.science/.well-known/pith/AKUOQVQN52P2DJ67E2FONQFX7Q/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:AKUOQVQN52P2DJ67E2FONQFX7Q","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":"76aff074b5659055ae86044c4066f2650493af99142ff8971d6ed9d78d439499","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-01-31T18:10:53Z","title_canon_sha256":"5d44f02811377d4cb6672388eeeb2f70ccf2255126fa1f5b1d4a7aede10fd668"},"schema_version":"1.0","source":{"id":"2501.19358","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2501.19358","created_at":"2026-07-05T11:14:03Z"},{"alias_kind":"arxiv_version","alias_value":"2501.19358v3","created_at":"2026-07-05T11:14:03Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.19358","created_at":"2026-07-05T11:14:03Z"},{"alias_kind":"pith_short_12","alias_value":"AKUOQVQN52P2","created_at":"2026-07-05T11:14:03Z"},{"alias_kind":"pith_short_16","alias_value":"AKUOQVQN52P2DJ67","created_at":"2026-07-05T11:14:03Z"},{"alias_kind":"pith_short_8","alias_value":"AKUOQVQN","created_at":"2026-07-05T11:14:03Z"}],"graph_snapshots":[{"event_id":"sha256:27efa2cd7141183de9b2fc96dc2a97f287ea4695b6efef9c012aa991632f178b","target":"graph","created_at":"2026-07-05T11:14:03Z","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/2501.19358/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"This work identifies the Energy Loss Phenomenon in Reinforcement Learning from Human Feedback (RLHF) and its connection to reward hacking. Specifically, energy loss in the final layer of a Large Language Model (LLM) gradually increases during the RL process, with an excessive increase in energy loss characterizing reward hacking. Beyond empirical analysis, we further provide a theoretical foundation by proving that, under mild conditions, the increased energy loss reduces the upper bound of contextual relevance in LLMs, which is a critical aspect of reward hacking as the reduced contextual rel","authors_text":"Dacheng Tao, Lefei Zhang, Liang Ding, Sen Zhang, Yuchun Miao, Yuqi Zhang","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-01-31T18:10:53Z","title":"The Energy Loss Phenomenon in RLHF: A New Perspective on Mitigating Reward Hacking"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.19358","kind":"arxiv","version":3},"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":"sha256:03029d4abd6d83321b86c882e347ae922a25c6ad5f208ea3f16b2832c42d3d93","target":"record","created_at":"2026-07-05T11:14:03Z","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":"76aff074b5659055ae86044c4066f2650493af99142ff8971d6ed9d78d439499","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-01-31T18:10:53Z","title_canon_sha256":"5d44f02811377d4cb6672388eeeb2f70ccf2255126fa1f5b1d4a7aede10fd668"},"schema_version":"1.0","source":{"id":"2501.19358","kind":"arxiv","version":3}},"canonical_sha256":"02a8e8560dee9fa1a7df268ae6c0b7fc0c650d2721f7bf85dec6e2d0747d48d0","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"02a8e8560dee9fa1a7df268ae6c0b7fc0c650d2721f7bf85dec6e2d0747d48d0","first_computed_at":"2026-07-05T11:14:03.018222Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:14:03.018222Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"JMWcqfnbkwmbsv3n6H6BfcdY4qTU+T4J1sZdc/NM+jCs/10H1xO2oz7HPRyMYmzR25P1Y+W78tSHF20o2ZfFCw==","signature_status":"signed_v1","signed_at":"2026-07-05T11:14:03.018776Z","signed_message":"canonical_sha256_bytes"},"source_id":"2501.19358","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:03029d4abd6d83321b86c882e347ae922a25c6ad5f208ea3f16b2832c42d3d93","sha256:27efa2cd7141183de9b2fc96dc2a97f287ea4695b6efef9c012aa991632f178b"],"state_sha256":"10be92ef94300501cddea7dd01e35f869bf091825c3bc52ad83c3300afbbc13d"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Aod8tUO12rd9LCLXUXhbVDE8zcr0Xm65nRSucnBaHpBsyEhvxkolVrpmEcsc1YN5G5c17CmHI5623y5qLcUbDQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-07T03:59:39.298833Z","bundle_sha256":"bb31ee81480b651a563b9dca81e259cf7a1069aa7f79a8f0d0eae9dc691ad025"}}