{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:6UZGADZMYHG6NVBEIAIFLUXH7X","short_pith_number":"pith:6UZGADZM","canonical_record":{"source":{"id":"2505.23556","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-05-29T15:33:39Z","cross_cats_sorted":[],"title_canon_sha256":"9fb52cffbc396e5ae8968e55ce022d4154bc71c629fdd3e0b3f544dead9e2803","abstract_canon_sha256":"f8977356cb7d094d15c1bd682b0c120ac401f76d4709be723c7fc30fadc4acc1"},"schema_version":"1.0"},"canonical_sha256":"f532600f2cc1cde6d424401055d2e7fddaa2530b4ebbce14d5729e9f4a077a3a","source":{"kind":"arxiv","id":"2505.23556","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2505.23556","created_at":"2026-07-05T11:12:03Z"},{"alias_kind":"arxiv_version","alias_value":"2505.23556v1","created_at":"2026-07-05T11:12:03Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.23556","created_at":"2026-07-05T11:12:03Z"},{"alias_kind":"pith_short_12","alias_value":"6UZGADZMYHG6","created_at":"2026-07-05T11:12:03Z"},{"alias_kind":"pith_short_16","alias_value":"6UZGADZMYHG6NVBE","created_at":"2026-07-05T11:12:03Z"},{"alias_kind":"pith_short_8","alias_value":"6UZGADZM","created_at":"2026-07-05T11:12:03Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:6UZGADZMYHG6NVBEIAIFLUXH7X","target":"record","payload":{"canonical_record":{"source":{"id":"2505.23556","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-05-29T15:33:39Z","cross_cats_sorted":[],"title_canon_sha256":"9fb52cffbc396e5ae8968e55ce022d4154bc71c629fdd3e0b3f544dead9e2803","abstract_canon_sha256":"f8977356cb7d094d15c1bd682b0c120ac401f76d4709be723c7fc30fadc4acc1"},"schema_version":"1.0"},"canonical_sha256":"f532600f2cc1cde6d424401055d2e7fddaa2530b4ebbce14d5729e9f4a077a3a","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:12:03.862859Z","signature_b64":"cQ4A/DlGLzv42WgjIpAnYlxDp+3IkXqozE3TTrjc7YKlA0O0VEwr+NMctCKtchnEefVId9K2gk6Z/PioOv1lDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f532600f2cc1cde6d424401055d2e7fddaa2530b4ebbce14d5729e9f4a077a3a","last_reissued_at":"2026-07-05T11:12:03.862380Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:12:03.862380Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2505.23556","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":"2026-07-05T11:12:03Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"uTsX2sxp65p8K/Q8mdes0I0xV5EBNuI+BxMXLDcwEKZyULg+WAc0Za1tup/lCndujyfPPKMUF8G+PZNgc7KXDA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-15T23:39:13.497377Z"},"content_sha256":"d77bd42571d34e8f7ac40c2b5a2f76bc06be3723587948c014ceb919382616a6","schema_version":"1.0","event_id":"sha256:d77bd42571d34e8f7ac40c2b5a2f76bc06be3723587948c014ceb919382616a6"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:6UZGADZMYHG6NVBEIAIFLUXH7X","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Understanding Refusal in Language Models with Sparse Autoencoders","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Clement Neo, Erik Cambria, Nirmalendu Prakash, Ranjan Satapathy, Roy Ka-Wei Lee, Wei Jie Yeo","submitted_at":"2025-05-29T15:33:39Z","abstract_excerpt":"Refusal is a key safety behavior in aligned language models, yet the internal mechanisms driving refusals remain opaque. In this work, we conduct a mechanistic study of refusal in instruction-tuned LLMs using sparse autoencoders to identify latent features that causally mediate refusal behaviors. We apply our method to two open-source chat models and intervene on refusal-related features to assess their influence on generation, validating their behavioral impact across multiple harmful datasets. This enables a fine-grained inspection of how refusal manifests at the activation level and address"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.23556","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/2505.23556/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":"2026-07-05T11:12:03Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"eWoYB9VhBpdx9Sc5uMY7PeBB/Q7vLnj14mXfSMXFN8SEa3YICceE9/wNnidfklS2TFL6vIAPY7V6Io4nIYqMAw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-15T23:39:13.497889Z"},"content_sha256":"1fd02f31d4ad285fa834a63a101166eb5f691a2e1144e1a3ed04f47b8e97ff56","schema_version":"1.0","event_id":"sha256:1fd02f31d4ad285fa834a63a101166eb5f691a2e1144e1a3ed04f47b8e97ff56"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/6UZGADZMYHG6NVBEIAIFLUXH7X/bundle.json","state_url":"https://pith.science/pith/6UZGADZMYHG6NVBEIAIFLUXH7X/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/6UZGADZMYHG6NVBEIAIFLUXH7X/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":"2026-08-15T23:39:13Z","links":{"resolver":"https://pith.science/pith/6UZGADZMYHG6NVBEIAIFLUXH7X","bundle":"https://pith.science/pith/6UZGADZMYHG6NVBEIAIFLUXH7X/bundle.json","state":"https://pith.science/pith/6UZGADZMYHG6NVBEIAIFLUXH7X/state.json","well_known_bundle":"https://pith.science/.well-known/pith/6UZGADZMYHG6NVBEIAIFLUXH7X/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:6UZGADZMYHG6NVBEIAIFLUXH7X","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":"f8977356cb7d094d15c1bd682b0c120ac401f76d4709be723c7fc30fadc4acc1","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-05-29T15:33:39Z","title_canon_sha256":"9fb52cffbc396e5ae8968e55ce022d4154bc71c629fdd3e0b3f544dead9e2803"},"schema_version":"1.0","source":{"id":"2505.23556","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2505.23556","created_at":"2026-07-05T11:12:03Z"},{"alias_kind":"arxiv_version","alias_value":"2505.23556v1","created_at":"2026-07-05T11:12:03Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.23556","created_at":"2026-07-05T11:12:03Z"},{"alias_kind":"pith_short_12","alias_value":"6UZGADZMYHG6","created_at":"2026-07-05T11:12:03Z"},{"alias_kind":"pith_short_16","alias_value":"6UZGADZMYHG6NVBE","created_at":"2026-07-05T11:12:03Z"},{"alias_kind":"pith_short_8","alias_value":"6UZGADZM","created_at":"2026-07-05T11:12:03Z"}],"graph_snapshots":[{"event_id":"sha256:1fd02f31d4ad285fa834a63a101166eb5f691a2e1144e1a3ed04f47b8e97ff56","target":"graph","created_at":"2026-07-05T11:12:03Z","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/2505.23556/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Refusal is a key safety behavior in aligned language models, yet the internal mechanisms driving refusals remain opaque. In this work, we conduct a mechanistic study of refusal in instruction-tuned LLMs using sparse autoencoders to identify latent features that causally mediate refusal behaviors. We apply our method to two open-source chat models and intervene on refusal-related features to assess their influence on generation, validating their behavioral impact across multiple harmful datasets. This enables a fine-grained inspection of how refusal manifests at the activation level and address","authors_text":"Clement Neo, Erik Cambria, Nirmalendu Prakash, Ranjan Satapathy, Roy Ka-Wei Lee, Wei Jie Yeo","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-05-29T15:33:39Z","title":"Understanding Refusal in Language Models with Sparse Autoencoders"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.23556","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:d77bd42571d34e8f7ac40c2b5a2f76bc06be3723587948c014ceb919382616a6","target":"record","created_at":"2026-07-05T11:12:03Z","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":"f8977356cb7d094d15c1bd682b0c120ac401f76d4709be723c7fc30fadc4acc1","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-05-29T15:33:39Z","title_canon_sha256":"9fb52cffbc396e5ae8968e55ce022d4154bc71c629fdd3e0b3f544dead9e2803"},"schema_version":"1.0","source":{"id":"2505.23556","kind":"arxiv","version":1}},"canonical_sha256":"f532600f2cc1cde6d424401055d2e7fddaa2530b4ebbce14d5729e9f4a077a3a","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"f532600f2cc1cde6d424401055d2e7fddaa2530b4ebbce14d5729e9f4a077a3a","first_computed_at":"2026-07-05T11:12:03.862380Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:12:03.862380Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"cQ4A/DlGLzv42WgjIpAnYlxDp+3IkXqozE3TTrjc7YKlA0O0VEwr+NMctCKtchnEefVId9K2gk6Z/PioOv1lDg==","signature_status":"signed_v1","signed_at":"2026-07-05T11:12:03.862859Z","signed_message":"canonical_sha256_bytes"},"source_id":"2505.23556","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:d77bd42571d34e8f7ac40c2b5a2f76bc06be3723587948c014ceb919382616a6","sha256:1fd02f31d4ad285fa834a63a101166eb5f691a2e1144e1a3ed04f47b8e97ff56"],"state_sha256":"e5657858c5666412eb2f97095024d2a393ee6ab90a4bdaaba9f7fa73081d2852"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"dP/Bc8VXDIS+peRQYfN4IcnLLvOX5Y0uTXT7tFLaOP0DgM0kx8rxr09Zb2/raH5faOrCJrc415uo1nBrVR7gDQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-15T23:39:13.502393Z","bundle_sha256":"e8d200c2a4b8fff6115edf98e7027f28acf046c9bad1b0f34b4406e616e546f0"}}