{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:FYUXIGPJADALVCJHAURQJIEGOG","short_pith_number":"pith:FYUXIGPJ","canonical_record":{"source":{"id":"2605.30487","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-28T19:05:53Z","cross_cats_sorted":[],"title_canon_sha256":"38380cd0b8e069e25110f6eae8b8ba96be0eed8b33be03889af4177bfa8c934d","abstract_canon_sha256":"fdac5d8be7e5aec69696dbc735847ace4e4a91fab6d0bc223c51102f6eff58af"},"schema_version":"1.0"},"canonical_sha256":"2e297419e900c0ba8927052304a08671b157de23e9c0ae462716670f3f1c08c4","source":{"kind":"arxiv","id":"2605.30487","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.30487","created_at":"2026-06-01T01:02:57Z"},{"alias_kind":"arxiv_version","alias_value":"2605.30487v1","created_at":"2026-06-01T01:02:57Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.30487","created_at":"2026-06-01T01:02:57Z"},{"alias_kind":"pith_short_12","alias_value":"FYUXIGPJADAL","created_at":"2026-06-01T01:02:57Z"},{"alias_kind":"pith_short_16","alias_value":"FYUXIGPJADALVCJH","created_at":"2026-06-01T01:02:57Z"},{"alias_kind":"pith_short_8","alias_value":"FYUXIGPJ","created_at":"2026-06-01T01:02:57Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:FYUXIGPJADALVCJHAURQJIEGOG","target":"record","payload":{"canonical_record":{"source":{"id":"2605.30487","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-28T19:05:53Z","cross_cats_sorted":[],"title_canon_sha256":"38380cd0b8e069e25110f6eae8b8ba96be0eed8b33be03889af4177bfa8c934d","abstract_canon_sha256":"fdac5d8be7e5aec69696dbc735847ace4e4a91fab6d0bc223c51102f6eff58af"},"schema_version":"1.0"},"canonical_sha256":"2e297419e900c0ba8927052304a08671b157de23e9c0ae462716670f3f1c08c4","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-01T01:02:57.024788Z","signature_b64":"NSm5nNr2OPF/ccgCoqeb4dXG14HDjreymiMH+o6XtSi0rfEOR883NHZ1i0y3Vu6CA3XJXqFJvtWvVDlDTxzlBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2e297419e900c0ba8927052304a08671b157de23e9c0ae462716670f3f1c08c4","last_reissued_at":"2026-06-01T01:02:57.023902Z","signature_status":"signed_v1","first_computed_at":"2026-06-01T01:02:57.023902Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2605.30487","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":"2026-06-01T01:02:57Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Uf2PUlEfRwJkrxe77DY8ziUnYlUq+iDY7cLbK2qe0wZalpAcqN8q7EuqW2IB6+ScGLC9VyDhJJc+tY7u5TQGAQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-04T14:23:09.890645Z"},"content_sha256":"2ef1c3488dcd7bbed3afcce1d4407fa7cccd1c041d4ef3102a0633da35163963","schema_version":"1.0","event_id":"sha256:2ef1c3488dcd7bbed3afcce1d4407fa7cccd1c041d4ef3102a0633da35163963"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:FYUXIGPJADALVCJHAURQJIEGOG","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Configurable Reward Model for Balanced Safety Alignment","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Akash Bharadwaj, Anqi Liu, Benjamin Van Durme, Li Chen, Manik Bhandari, Mayur Srungarapu, Mehran Khodabandeh, Zhengping Jiang","submitted_at":"2026-05-28T19:05:53Z","abstract_excerpt":"Aligning large language models (LLMs) to heterogeneous and rapidly evolving safety requirements remains a critical challenge. Existing instruction-tuned LLMs and standalone safety classifiers often fail to generalize to new safety configurations, motivating the need for Reward Models (RMs) that are explicitly configurable to changing specifications. We introduce the Configurable Safety Reward Model (CSRM), which is jointly optimized for calibrated safety compliance and reward modeling. Our approach is supported by configuration-targeted data augmentation that enforces instruction adherence whi"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.30487","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/2605.30487/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":"2026-06-01T01:02:57Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"mtMNAFIRHZNP95PVms97FUIwqQWQgoHODy15D67l8D8hfuAaQAYFisZoR4Pr0EQJEN9hfisYvifPNmdKoeWgDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-04T14:23:09.891036Z"},"content_sha256":"93a7038ddc0545b325b46829421c928c819bf3578c5ed77b2bef8e545c4d9740","schema_version":"1.0","event_id":"sha256:93a7038ddc0545b325b46829421c928c819bf3578c5ed77b2bef8e545c4d9740"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/FYUXIGPJADALVCJHAURQJIEGOG/bundle.json","state_url":"https://pith.science/pith/FYUXIGPJADALVCJHAURQJIEGOG/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/FYUXIGPJADALVCJHAURQJIEGOG/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":"2026-06-04T14:23:09Z","links":{"resolver":"https://pith.science/pith/FYUXIGPJADALVCJHAURQJIEGOG","bundle":"https://pith.science/pith/FYUXIGPJADALVCJHAURQJIEGOG/bundle.json","state":"https://pith.science/pith/FYUXIGPJADALVCJHAURQJIEGOG/state.json","well_known_bundle":"https://pith.science/.well-known/pith/FYUXIGPJADALVCJHAURQJIEGOG/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:FYUXIGPJADALVCJHAURQJIEGOG","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":"fdac5d8be7e5aec69696dbc735847ace4e4a91fab6d0bc223c51102f6eff58af","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-28T19:05:53Z","title_canon_sha256":"38380cd0b8e069e25110f6eae8b8ba96be0eed8b33be03889af4177bfa8c934d"},"schema_version":"1.0","source":{"id":"2605.30487","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.30487","created_at":"2026-06-01T01:02:57Z"},{"alias_kind":"arxiv_version","alias_value":"2605.30487v1","created_at":"2026-06-01T01:02:57Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.30487","created_at":"2026-06-01T01:02:57Z"},{"alias_kind":"pith_short_12","alias_value":"FYUXIGPJADAL","created_at":"2026-06-01T01:02:57Z"},{"alias_kind":"pith_short_16","alias_value":"FYUXIGPJADALVCJH","created_at":"2026-06-01T01:02:57Z"},{"alias_kind":"pith_short_8","alias_value":"FYUXIGPJ","created_at":"2026-06-01T01:02:57Z"}],"graph_snapshots":[{"event_id":"sha256:93a7038ddc0545b325b46829421c928c819bf3578c5ed77b2bef8e545c4d9740","target":"graph","created_at":"2026-06-01T01:02:57Z","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/2605.30487/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Aligning large language models (LLMs) to heterogeneous and rapidly evolving safety requirements remains a critical challenge. Existing instruction-tuned LLMs and standalone safety classifiers often fail to generalize to new safety configurations, motivating the need for Reward Models (RMs) that are explicitly configurable to changing specifications. We introduce the Configurable Safety Reward Model (CSRM), which is jointly optimized for calibrated safety compliance and reward modeling. Our approach is supported by configuration-targeted data augmentation that enforces instruction adherence whi","authors_text":"Akash Bharadwaj, Anqi Liu, Benjamin Van Durme, Li Chen, Manik Bhandari, Mayur Srungarapu, Mehran Khodabandeh, Zhengping Jiang","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-28T19:05:53Z","title":"Configurable Reward Model for Balanced Safety Alignment"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.30487","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":"sha256:2ef1c3488dcd7bbed3afcce1d4407fa7cccd1c041d4ef3102a0633da35163963","target":"record","created_at":"2026-06-01T01:02:57Z","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":"fdac5d8be7e5aec69696dbc735847ace4e4a91fab6d0bc223c51102f6eff58af","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-28T19:05:53Z","title_canon_sha256":"38380cd0b8e069e25110f6eae8b8ba96be0eed8b33be03889af4177bfa8c934d"},"schema_version":"1.0","source":{"id":"2605.30487","kind":"arxiv","version":1}},"canonical_sha256":"2e297419e900c0ba8927052304a08671b157de23e9c0ae462716670f3f1c08c4","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"2e297419e900c0ba8927052304a08671b157de23e9c0ae462716670f3f1c08c4","first_computed_at":"2026-06-01T01:02:57.023902Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-06-01T01:02:57.023902Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"NSm5nNr2OPF/ccgCoqeb4dXG14HDjreymiMH+o6XtSi0rfEOR883NHZ1i0y3Vu6CA3XJXqFJvtWvVDlDTxzlBw==","signature_status":"signed_v1","signed_at":"2026-06-01T01:02:57.024788Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.30487","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:2ef1c3488dcd7bbed3afcce1d4407fa7cccd1c041d4ef3102a0633da35163963","sha256:93a7038ddc0545b325b46829421c928c819bf3578c5ed77b2bef8e545c4d9740"],"state_sha256":"9b45fee031468f74115ff18e169f341ee29dae50fe649ccc5c6b394c379846eb"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"tAYXb+pcIDBlDp6wbfnbTWs6YoGBu2HdfNCa/du6Jc3twJaWMvuwwJb88BZH3qDzVAM7swL7yT39L5SMp5alCw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-04T14:23:09.893298Z","bundle_sha256":"0985c93015adb963bd4a8b0736b0ab64bd571b48cc2f5b43dcac0fbfb05dc59d"}}