{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:DETCQ2VL4XUORD7WOVMQE4IAAB","short_pith_number":"pith:DETCQ2VL","canonical_record":{"source":{"id":"2605.21318","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-20T15:47:26Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"be4860dcd92a8f1ad63e41791bd7bc8f7f8f33f0ec868ad216be646f1c653677","abstract_canon_sha256":"c374bba1091653ea07d3a423ea9c8e61d4458ebcd6cbbfb7ad9b1392e0195844"},"schema_version":"1.0"},"canonical_sha256":"1926286aabe5e8e88ff6755902710000521e4f7e475a8985cb809efdb5b6b5b2","source":{"kind":"arxiv","id":"2605.21318","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.21318","created_at":"2026-05-21T02:05:28Z"},{"alias_kind":"arxiv_version","alias_value":"2605.21318v1","created_at":"2026-05-21T02:05:28Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.21318","created_at":"2026-05-21T02:05:28Z"},{"alias_kind":"pith_short_12","alias_value":"DETCQ2VL4XUO","created_at":"2026-05-21T02:05:28Z"},{"alias_kind":"pith_short_16","alias_value":"DETCQ2VL4XUORD7W","created_at":"2026-05-21T02:05:28Z"},{"alias_kind":"pith_short_8","alias_value":"DETCQ2VL","created_at":"2026-05-21T02:05:28Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:DETCQ2VL4XUORD7WOVMQE4IAAB","target":"record","payload":{"canonical_record":{"source":{"id":"2605.21318","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-20T15:47:26Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"be4860dcd92a8f1ad63e41791bd7bc8f7f8f33f0ec868ad216be646f1c653677","abstract_canon_sha256":"c374bba1091653ea07d3a423ea9c8e61d4458ebcd6cbbfb7ad9b1392e0195844"},"schema_version":"1.0"},"canonical_sha256":"1926286aabe5e8e88ff6755902710000521e4f7e475a8985cb809efdb5b6b5b2","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-21T02:05:28.967440Z","signature_b64":"A+nDNTd0bqSHaykJeU745dg/Bt/gzOBNL0To/aF6hqpDt15F8ajWqd2l6pAe5OFkIv1uNwDje47EwGU59c2zBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1926286aabe5e8e88ff6755902710000521e4f7e475a8985cb809efdb5b6b5b2","last_reissued_at":"2026-05-21T02:05:28.966606Z","signature_status":"signed_v1","first_computed_at":"2026-05-21T02:05:28.966606Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2605.21318","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":"2026-05-21T02:05:28Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"dLhtWmU1xINWtfforh6OPwDrer3vVFTjjDAjyMn55+YU1k5JgJR5d5viibIlPS39UrScfXnC2rzh9t4XQLlMDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-24T12:22:12.699802Z"},"content_sha256":"bcfaae552d72cb8e16e66e63a46c53eeb5c595c9b4d68a4ccd64b0b5a2d4dbe7","schema_version":"1.0","event_id":"sha256:bcfaae552d72cb8e16e66e63a46c53eeb5c595c9b4d68a4ccd64b0b5a2d4dbe7"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:DETCQ2VL4XUORD7WOVMQE4IAAB","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"TextReg: Mitigating Prompt Distributional Overfitting via Regularized Text-Space Optimization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"B. Aditya Prakash, Haibo Jin, Haohan Wang, Lucheng Fu, Ye Yu, Yiqiao Jin, Yiyang Wang","submitted_at":"2026-05-20T15:47:26Z","abstract_excerpt":"Large language models (LLMs) are highly sensitive to the prompts used to specify task objectives and behavioral constraints. Many recent prompt optimization methods iteratively rewrite prompts using LLM-generated feedback, but the resulting prompts often become longer, accumulate narrow sample-specific rules, and generalize poorly beyond the training distribution. We study this failure mode as prompt distributional overfitting and argue that it reflects a lack of representation control in discrete text-space optimization. We formalize this view through representational inefficiency, a dual-fac"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.21318","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/2605.21318/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":"2026-05-21T02:05:28Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"QdPW6DWixv3PaMoZ1qAemf/Wz8gQCKBOOsjego8r04XRb/4Sdd1te4ppa91oqvsvxnGSUXiHUqWd9MIobNGkAA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-24T12:22:12.700189Z"},"content_sha256":"ca7f4bfd1bfb82bcb0eab03ec931e70350e4c1aad1e7dbcc02e7f490a8830d23","schema_version":"1.0","event_id":"sha256:ca7f4bfd1bfb82bcb0eab03ec931e70350e4c1aad1e7dbcc02e7f490a8830d23"},{"event_type":"integrity_finding","subject_pith_number":"pith:2026:DETCQ2VL4XUORD7WOVMQE4IAAB","target":"integrity","payload":{"note":"Citing paper attributes a specific factual claim to reference [9], which resolves to arXiv:2406.07496. The claim's distinctive tokens have only 13% overlap with any chunk of the cited paper's stored text (threshold for unsupported is 15%). The attribution could not be verified against the cited work.","snippet":"this train–generalization mismatch, while DLPO [30] mitigates overfitting through static simplification\ninstructions, and REMO [31] uses external memory and meta-reflection.","arxiv_id":"2605.21318","detector":"citation_quote_validity","evidence":{"ref_index":9,"claim_text":"this train–generalization mismatch, while DLPO [30] mitigates overfitting through static simplification\ninstructions, and REMO [31] uses external memory and meta-reflection.","claim_offset":9032,"verdict_class":"threshold_with_margin","chunks_checked":14,"cited_arxiv_id":"2406.07496","best_overlap_score":0.133,"supported_threshold":0.5,"best_overlap_chunk_id":"e80bff58-95d6-4cac-9303-c3c3fb499540","unsupported_threshold":0.15},"severity":"advisory","ref_index":9,"audited_at":"2026-05-21T07:49:58.589743Z","event_type":"pith.integrity.v1","detected_doi":null,"detector_url":"https://pith.science/pith-integrity-protocol#citation_quote_validity","external_url":null,"finding_type":"unsupported_attribution","evidence_hash":"fa9158727af4a86750228b9f2b6d2287b332bd43ca72485ff00751f5c00df410","paper_version":1,"verdict_class":"threshold_with_margin","resolved_title":null,"detector_version":"0.1.0","detected_arxiv_id":"2406.07496","integrity_event_id":5872,"payload_sha256":"4ff9eb0db5d127ee96a45664b0cf57833abe41d30b4b4b9f3aec85600fc43c04","signature_b64":"qEj2ZpfdnXaVsdhunhI17hSOcHineGagsPVpook0ubGrJalpFpJ1D4c/vqvt0W40bYZdSVEsEBhh5koWi26WDQ==","signing_key_id":"pith-v1-2026-05"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-21T07:54:12Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"nMu81altujIv8LmUMqMbY38R2YM/F+8mrkJjeacudl5V8s1YDu/heh2IdovggHtImF88YZm1dxJA6kz3ZggoAA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-24T12:22:12.701629Z"},"content_sha256":"5c68edf252fc8e0292135f0b182fcc67bdaa50d3780e16928378d895c576224f","schema_version":"1.0","event_id":"sha256:5c68edf252fc8e0292135f0b182fcc67bdaa50d3780e16928378d895c576224f"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/DETCQ2VL4XUORD7WOVMQE4IAAB/bundle.json","state_url":"https://pith.science/pith/DETCQ2VL4XUORD7WOVMQE4IAAB/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/DETCQ2VL4XUORD7WOVMQE4IAAB/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":"2026-05-24T12:22:12Z","links":{"resolver":"https://pith.science/pith/DETCQ2VL4XUORD7WOVMQE4IAAB","bundle":"https://pith.science/pith/DETCQ2VL4XUORD7WOVMQE4IAAB/bundle.json","state":"https://pith.science/pith/DETCQ2VL4XUORD7WOVMQE4IAAB/state.json","well_known_bundle":"https://pith.science/.well-known/pith/DETCQ2VL4XUORD7WOVMQE4IAAB/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:DETCQ2VL4XUORD7WOVMQE4IAAB","merge_version":"pith-open-graph-merge-v1","event_count":3,"valid_event_count":3,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"c374bba1091653ea07d3a423ea9c8e61d4458ebcd6cbbfb7ad9b1392e0195844","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-20T15:47:26Z","title_canon_sha256":"be4860dcd92a8f1ad63e41791bd7bc8f7f8f33f0ec868ad216be646f1c653677"},"schema_version":"1.0","source":{"id":"2605.21318","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.21318","created_at":"2026-05-21T02:05:28Z"},{"alias_kind":"arxiv_version","alias_value":"2605.21318v1","created_at":"2026-05-21T02:05:28Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.21318","created_at":"2026-05-21T02:05:28Z"},{"alias_kind":"pith_short_12","alias_value":"DETCQ2VL4XUO","created_at":"2026-05-21T02:05:28Z"},{"alias_kind":"pith_short_16","alias_value":"DETCQ2VL4XUORD7W","created_at":"2026-05-21T02:05:28Z"},{"alias_kind":"pith_short_8","alias_value":"DETCQ2VL","created_at":"2026-05-21T02:05:28Z"}],"graph_snapshots":[{"event_id":"sha256:ca7f4bfd1bfb82bcb0eab03ec931e70350e4c1aad1e7dbcc02e7f490a8830d23","target":"graph","created_at":"2026-05-21T02:05:28Z","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/2605.21318/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Large language models (LLMs) are highly sensitive to the prompts used to specify task objectives and behavioral constraints. Many recent prompt optimization methods iteratively rewrite prompts using LLM-generated feedback, but the resulting prompts often become longer, accumulate narrow sample-specific rules, and generalize poorly beyond the training distribution. We study this failure mode as prompt distributional overfitting and argue that it reflects a lack of representation control in discrete text-space optimization. We formalize this view through representational inefficiency, a dual-fac","authors_text":"B. Aditya Prakash, Haibo Jin, Haohan Wang, Lucheng Fu, Ye Yu, Yiqiao Jin, Yiyang Wang","cross_cats":["cs.AI","cs.LG"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-20T15:47:26Z","title":"TextReg: Mitigating Prompt Distributional Overfitting via Regularized Text-Space Optimization"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.21318","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:bcfaae552d72cb8e16e66e63a46c53eeb5c595c9b4d68a4ccd64b0b5a2d4dbe7","target":"record","created_at":"2026-05-21T02:05:28Z","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":"c374bba1091653ea07d3a423ea9c8e61d4458ebcd6cbbfb7ad9b1392e0195844","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-20T15:47:26Z","title_canon_sha256":"be4860dcd92a8f1ad63e41791bd7bc8f7f8f33f0ec868ad216be646f1c653677"},"schema_version":"1.0","source":{"id":"2605.21318","kind":"arxiv","version":1}},"canonical_sha256":"1926286aabe5e8e88ff6755902710000521e4f7e475a8985cb809efdb5b6b5b2","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"1926286aabe5e8e88ff6755902710000521e4f7e475a8985cb809efdb5b6b5b2","first_computed_at":"2026-05-21T02:05:28.966606Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-21T02:05:28.966606Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"A+nDNTd0bqSHaykJeU745dg/Bt/gzOBNL0To/aF6hqpDt15F8ajWqd2l6pAe5OFkIv1uNwDje47EwGU59c2zBw==","signature_status":"signed_v1","signed_at":"2026-05-21T02:05:28.967440Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.21318","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:bcfaae552d72cb8e16e66e63a46c53eeb5c595c9b4d68a4ccd64b0b5a2d4dbe7","sha256:ca7f4bfd1bfb82bcb0eab03ec931e70350e4c1aad1e7dbcc02e7f490a8830d23","sha256:5c68edf252fc8e0292135f0b182fcc67bdaa50d3780e16928378d895c576224f"],"state_sha256":"c3ab4be7342ff6a557a2d74ff9428dd159bd45e2a954ccf172fdb8fda301a167"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"QZpvN9vOeB95qKFJf9mOXgyLAmT476FDGz//H0G2YWUyyNPC1CTtPgocMXQVKeyy4ywGbRwwK/tJ6btN4o3cBA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-24T12:22:12.704795Z","bundle_sha256":"ad3db65f33cf3ccd890a203e905173633d8b0e0b1de2c2c9f63a8812596a822d"}}