{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:NIIFT4C776SLOE3MYZD74EO6FD","short_pith_number":"pith:NIIFT4C7","canonical_record":{"source":{"id":"2605.23970","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-13T07:00:16Z","cross_cats_sorted":[],"title_canon_sha256":"04c713edb880d3bf461ead4125ae42718f79eae0dde0ad1f143fb71a6c15a509","abstract_canon_sha256":"d1637c947c9ff2de692acb9497ae5d731b0a2139190127f38fab7b2348f31920"},"schema_version":"1.0"},"canonical_sha256":"6a1059f05fffa4b7136cc647fe11de28db8f9795c95416e224d7c6778c4f50b9","source":{"kind":"arxiv","id":"2605.23970","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.23970","created_at":"2026-05-26T00:02:24Z"},{"alias_kind":"arxiv_version","alias_value":"2605.23970v1","created_at":"2026-05-26T00:02:24Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.23970","created_at":"2026-05-26T00:02:24Z"},{"alias_kind":"pith_short_12","alias_value":"NIIFT4C776SL","created_at":"2026-05-26T00:02:24Z"},{"alias_kind":"pith_short_16","alias_value":"NIIFT4C776SLOE3M","created_at":"2026-05-26T00:02:24Z"},{"alias_kind":"pith_short_8","alias_value":"NIIFT4C7","created_at":"2026-05-26T00:02:24Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:NIIFT4C776SLOE3MYZD74EO6FD","target":"record","payload":{"canonical_record":{"source":{"id":"2605.23970","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-13T07:00:16Z","cross_cats_sorted":[],"title_canon_sha256":"04c713edb880d3bf461ead4125ae42718f79eae0dde0ad1f143fb71a6c15a509","abstract_canon_sha256":"d1637c947c9ff2de692acb9497ae5d731b0a2139190127f38fab7b2348f31920"},"schema_version":"1.0"},"canonical_sha256":"6a1059f05fffa4b7136cc647fe11de28db8f9795c95416e224d7c6778c4f50b9","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-26T00:02:24.503164Z","signature_b64":"bE+fFEzbNTo9drilIztHhVqgDg3XeM632HDNEahWN7rRfVsHu3lPLstKQP90AjhX6SDnXFKyPDmEezlAu5kBCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6a1059f05fffa4b7136cc647fe11de28db8f9795c95416e224d7c6778c4f50b9","last_reissued_at":"2026-05-26T00:02:24.502387Z","signature_status":"signed_v1","first_computed_at":"2026-05-26T00:02:24.502387Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2605.23970","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":"2026-05-26T00:02:24Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"bEGPbrnJGOVpko04hzPzGsewFIGAF4LsLSg/x+XEbTQF84ny6/s1i8Pe9+QDXazSjggaefoz1Fms1Ven9chnCw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-31T19:43:17.789231Z"},"content_sha256":"3c66dbe02f1d3cc26369a864524d86d1db2962587d179641d854c0564f30bd7b","schema_version":"1.0","event_id":"sha256:3c66dbe02f1d3cc26369a864524d86d1db2962587d179641d854c0564f30bd7b"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:NIIFT4C776SLOE3MYZD74EO6FD","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Faithful or Fabricated? A Causal Framework for Rationalization Bias in LLM Judges","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Abhishek Kumar, Carsten Maple, Riya Tapwal","submitted_at":"2026-05-13T07:00:16Z","abstract_excerpt":"Large language models (LLMs) are increasingly used as automatic judges for summarization and dialogue evaluation. Prior work has documented biases such as position, verbosity, and style preferences, but largely focuses on outcomes, leaving judge explanations underexplored. We instead ask whether LLM judges are cue-invariant, i.e., whether their rankings and explanations remain stable when non-evidential cues are perturbed while holding the underlying texts fixed. We introduce a suite of cue interventions (Blind, Truth, Flip, Placebo, Reveal-After) and tie-aware metrics that quantify outcome an"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.23970","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/2605.23970/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":"2026-05-26T00:02:24Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"xOUwNOyX+jdlgw5T25mTrXPvdfLZ56+j0OTGlwlTWq4nlYhhs2+PZyR87+ye7SLMEJvrG7h1C4XVDoUi51frCw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-31T19:43:17.790070Z"},"content_sha256":"9cf13d24389d12afdbbe41f13cebc47983a57ec69ae76b4cf6071c828f378e05","schema_version":"1.0","event_id":"sha256:9cf13d24389d12afdbbe41f13cebc47983a57ec69ae76b4cf6071c828f378e05"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/NIIFT4C776SLOE3MYZD74EO6FD/bundle.json","state_url":"https://pith.science/pith/NIIFT4C776SLOE3MYZD74EO6FD/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/NIIFT4C776SLOE3MYZD74EO6FD/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":"2026-05-31T19:43:17Z","links":{"resolver":"https://pith.science/pith/NIIFT4C776SLOE3MYZD74EO6FD","bundle":"https://pith.science/pith/NIIFT4C776SLOE3MYZD74EO6FD/bundle.json","state":"https://pith.science/pith/NIIFT4C776SLOE3MYZD74EO6FD/state.json","well_known_bundle":"https://pith.science/.well-known/pith/NIIFT4C776SLOE3MYZD74EO6FD/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:NIIFT4C776SLOE3MYZD74EO6FD","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":"d1637c947c9ff2de692acb9497ae5d731b0a2139190127f38fab7b2348f31920","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-13T07:00:16Z","title_canon_sha256":"04c713edb880d3bf461ead4125ae42718f79eae0dde0ad1f143fb71a6c15a509"},"schema_version":"1.0","source":{"id":"2605.23970","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.23970","created_at":"2026-05-26T00:02:24Z"},{"alias_kind":"arxiv_version","alias_value":"2605.23970v1","created_at":"2026-05-26T00:02:24Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.23970","created_at":"2026-05-26T00:02:24Z"},{"alias_kind":"pith_short_12","alias_value":"NIIFT4C776SL","created_at":"2026-05-26T00:02:24Z"},{"alias_kind":"pith_short_16","alias_value":"NIIFT4C776SLOE3M","created_at":"2026-05-26T00:02:24Z"},{"alias_kind":"pith_short_8","alias_value":"NIIFT4C7","created_at":"2026-05-26T00:02:24Z"}],"graph_snapshots":[{"event_id":"sha256:9cf13d24389d12afdbbe41f13cebc47983a57ec69ae76b4cf6071c828f378e05","target":"graph","created_at":"2026-05-26T00:02:24Z","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/2605.23970/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Large language models (LLMs) are increasingly used as automatic judges for summarization and dialogue evaluation. Prior work has documented biases such as position, verbosity, and style preferences, but largely focuses on outcomes, leaving judge explanations underexplored. We instead ask whether LLM judges are cue-invariant, i.e., whether their rankings and explanations remain stable when non-evidential cues are perturbed while holding the underlying texts fixed. We introduce a suite of cue interventions (Blind, Truth, Flip, Placebo, Reveal-After) and tie-aware metrics that quantify outcome an","authors_text":"Abhishek Kumar, Carsten Maple, Riya Tapwal","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-13T07:00:16Z","title":"Faithful or Fabricated? A Causal Framework for Rationalization Bias in LLM Judges"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.23970","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":"sha256:3c66dbe02f1d3cc26369a864524d86d1db2962587d179641d854c0564f30bd7b","target":"record","created_at":"2026-05-26T00:02:24Z","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":"d1637c947c9ff2de692acb9497ae5d731b0a2139190127f38fab7b2348f31920","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-13T07:00:16Z","title_canon_sha256":"04c713edb880d3bf461ead4125ae42718f79eae0dde0ad1f143fb71a6c15a509"},"schema_version":"1.0","source":{"id":"2605.23970","kind":"arxiv","version":1}},"canonical_sha256":"6a1059f05fffa4b7136cc647fe11de28db8f9795c95416e224d7c6778c4f50b9","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"6a1059f05fffa4b7136cc647fe11de28db8f9795c95416e224d7c6778c4f50b9","first_computed_at":"2026-05-26T00:02:24.502387Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-26T00:02:24.502387Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"bE+fFEzbNTo9drilIztHhVqgDg3XeM632HDNEahWN7rRfVsHu3lPLstKQP90AjhX6SDnXFKyPDmEezlAu5kBCQ==","signature_status":"signed_v1","signed_at":"2026-05-26T00:02:24.503164Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.23970","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:3c66dbe02f1d3cc26369a864524d86d1db2962587d179641d854c0564f30bd7b","sha256:9cf13d24389d12afdbbe41f13cebc47983a57ec69ae76b4cf6071c828f378e05"],"state_sha256":"06a4736589da42edb3c01a330faeb2deaacda5deab0b836836f8c9cc2d26dfbe"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"4M273/ITNGPS9IApMM27bhdV281RgLG2kLkLPXNgMPEc/qL/eRgJGRd7dIQeqBqpp/jZgBJkeVQyHIqP5VBWBQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-31T19:43:17.794109Z","bundle_sha256":"135ea681ff10337928d01b3f30d668a8924b5bd29bbd1e702f4c6310412226ee"}}