{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:IFGXB6KDPQGJRHVEBZFZ23OU2J","short_pith_number":"pith:IFGXB6KD","canonical_record":{"source":{"id":"2605.22817","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-21T17:59:26Z","cross_cats_sorted":["cs.AI","cs.CL","cs.NE"],"title_canon_sha256":"8c9e6f0d44355a96cbaa52663e5e25cdc827d4e9666d21ed80d6e497be03dd25","abstract_canon_sha256":"0e89ae17f489b27004868273bdcff22eaf6b8025b58f3176a58c9deb5564eb0f"},"schema_version":"1.0"},"canonical_sha256":"414d70f9437c0c989ea40e4b9d6dd4d263ee05bac92ab2688fe9fad14ede34e5","source":{"kind":"arxiv","id":"2605.22817","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.22817","created_at":"2026-05-22T02:04:56Z"},{"alias_kind":"arxiv_version","alias_value":"2605.22817v1","created_at":"2026-05-22T02:04:56Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.22817","created_at":"2026-05-22T02:04:56Z"},{"alias_kind":"pith_short_12","alias_value":"IFGXB6KDPQGJ","created_at":"2026-05-22T02:04:56Z"},{"alias_kind":"pith_short_16","alias_value":"IFGXB6KDPQGJRHVE","created_at":"2026-05-22T02:04:56Z"},{"alias_kind":"pith_short_8","alias_value":"IFGXB6KD","created_at":"2026-05-22T02:04:56Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:IFGXB6KDPQGJRHVEBZFZ23OU2J","target":"record","payload":{"canonical_record":{"source":{"id":"2605.22817","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-21T17:59:26Z","cross_cats_sorted":["cs.AI","cs.CL","cs.NE"],"title_canon_sha256":"8c9e6f0d44355a96cbaa52663e5e25cdc827d4e9666d21ed80d6e497be03dd25","abstract_canon_sha256":"0e89ae17f489b27004868273bdcff22eaf6b8025b58f3176a58c9deb5564eb0f"},"schema_version":"1.0"},"canonical_sha256":"414d70f9437c0c989ea40e4b9d6dd4d263ee05bac92ab2688fe9fad14ede34e5","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-22T02:04:56.206448Z","signature_b64":"qX8DQTrChHI1GIGruVTFeR7c8UEpF2b1pwPmJSWe0yh59jZIQHVMahUJAKyhadSPZ2gKXZZn6RpHZiTTn5DbBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"414d70f9437c0c989ea40e4b9d6dd4d263ee05bac92ab2688fe9fad14ede34e5","last_reissued_at":"2026-05-22T02:04:56.205978Z","signature_status":"signed_v1","first_computed_at":"2026-05-22T02:04:56.205978Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2605.22817","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":"2026-05-22T02:04:56Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"nz/4MePqbMpjMvA1xRi7erUhGQGnuYEUxg7QcUUxgpyWHRxSc2zvgU7EUFaJ5eMOAF+oYOAg9RSK+KdmNvLcAQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-26T11:10:49.390626Z"},"content_sha256":"67d014b285ab832f8adff12a98bb1d9a5b2c47acf9d754ff4e0269eeddabd1b7","schema_version":"1.0","event_id":"sha256:67d014b285ab832f8adff12a98bb1d9a5b2c47acf9d754ff4e0269eeddabd1b7"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:IFGXB6KDPQGJRHVEBZFZ23OU2J","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Vector Policy Optimization: Training for Diversity Improves Test-Time Search","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.NE"],"primary_cat":"cs.LG","authors_text":"Akarsh Kumar, Idan Shenfeld, Isha Puri, Mehul Damani, Omar Khattab, Pulkit Agrawal, Ryan Bahlous-Boldi, Sebastian Risi, Zhang-Wei Hong","submitted_at":"2026-05-21T17:59:26Z","abstract_excerpt":"Language models must now generalize out of the box to novel environments and work inside inference-scaling search procedures, such as AlphaEvolve, that select rollouts with a variety of task-specific reward functions. Unfortunately, the standard paradigm of LLM post-training optimizes a pre-specified scalar reward, often leading current LLMs to produce low-entropy response distributions and thus to struggle at displaying the diversity that inference-time search will require. We propose Vector Policy Optimization (VPO), an RL algorithm that explicitly trains policies to anticipate diverse downs"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.22817","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/2605.22817/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":"2026-05-22T02:04:56Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"8vrVa0HOOiw0sJFROtrlQcK1k1TPkDJ8et/e9iKdZnBgWIIEo2hv+YdBfz4rIEJHKGxa/k7CWOxhYAsR70f9Dg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-26T11:10:49.391199Z"},"content_sha256":"fcb0b1ca10bc36cbce43075484de5d2519bf38f298bf6f68bfaf37bef1eeaff4","schema_version":"1.0","event_id":"sha256:fcb0b1ca10bc36cbce43075484de5d2519bf38f298bf6f68bfaf37bef1eeaff4"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/IFGXB6KDPQGJRHVEBZFZ23OU2J/bundle.json","state_url":"https://pith.science/pith/IFGXB6KDPQGJRHVEBZFZ23OU2J/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/IFGXB6KDPQGJRHVEBZFZ23OU2J/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":"2026-05-26T11:10:49Z","links":{"resolver":"https://pith.science/pith/IFGXB6KDPQGJRHVEBZFZ23OU2J","bundle":"https://pith.science/pith/IFGXB6KDPQGJRHVEBZFZ23OU2J/bundle.json","state":"https://pith.science/pith/IFGXB6KDPQGJRHVEBZFZ23OU2J/state.json","well_known_bundle":"https://pith.science/.well-known/pith/IFGXB6KDPQGJRHVEBZFZ23OU2J/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:IFGXB6KDPQGJRHVEBZFZ23OU2J","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":"0e89ae17f489b27004868273bdcff22eaf6b8025b58f3176a58c9deb5564eb0f","cross_cats_sorted":["cs.AI","cs.CL","cs.NE"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-21T17:59:26Z","title_canon_sha256":"8c9e6f0d44355a96cbaa52663e5e25cdc827d4e9666d21ed80d6e497be03dd25"},"schema_version":"1.0","source":{"id":"2605.22817","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.22817","created_at":"2026-05-22T02:04:56Z"},{"alias_kind":"arxiv_version","alias_value":"2605.22817v1","created_at":"2026-05-22T02:04:56Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.22817","created_at":"2026-05-22T02:04:56Z"},{"alias_kind":"pith_short_12","alias_value":"IFGXB6KDPQGJ","created_at":"2026-05-22T02:04:56Z"},{"alias_kind":"pith_short_16","alias_value":"IFGXB6KDPQGJRHVE","created_at":"2026-05-22T02:04:56Z"},{"alias_kind":"pith_short_8","alias_value":"IFGXB6KD","created_at":"2026-05-22T02:04:56Z"}],"graph_snapshots":[{"event_id":"sha256:fcb0b1ca10bc36cbce43075484de5d2519bf38f298bf6f68bfaf37bef1eeaff4","target":"graph","created_at":"2026-05-22T02:04:56Z","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/2605.22817/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Language models must now generalize out of the box to novel environments and work inside inference-scaling search procedures, such as AlphaEvolve, that select rollouts with a variety of task-specific reward functions. Unfortunately, the standard paradigm of LLM post-training optimizes a pre-specified scalar reward, often leading current LLMs to produce low-entropy response distributions and thus to struggle at displaying the diversity that inference-time search will require. We propose Vector Policy Optimization (VPO), an RL algorithm that explicitly trains policies to anticipate diverse downs","authors_text":"Akarsh Kumar, Idan Shenfeld, Isha Puri, Mehul Damani, Omar Khattab, Pulkit Agrawal, Ryan Bahlous-Boldi, Sebastian Risi, Zhang-Wei Hong","cross_cats":["cs.AI","cs.CL","cs.NE"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-21T17:59:26Z","title":"Vector Policy Optimization: Training for Diversity Improves Test-Time Search"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.22817","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":"sha256:67d014b285ab832f8adff12a98bb1d9a5b2c47acf9d754ff4e0269eeddabd1b7","target":"record","created_at":"2026-05-22T02:04:56Z","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":"0e89ae17f489b27004868273bdcff22eaf6b8025b58f3176a58c9deb5564eb0f","cross_cats_sorted":["cs.AI","cs.CL","cs.NE"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-21T17:59:26Z","title_canon_sha256":"8c9e6f0d44355a96cbaa52663e5e25cdc827d4e9666d21ed80d6e497be03dd25"},"schema_version":"1.0","source":{"id":"2605.22817","kind":"arxiv","version":1}},"canonical_sha256":"414d70f9437c0c989ea40e4b9d6dd4d263ee05bac92ab2688fe9fad14ede34e5","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"414d70f9437c0c989ea40e4b9d6dd4d263ee05bac92ab2688fe9fad14ede34e5","first_computed_at":"2026-05-22T02:04:56.205978Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-22T02:04:56.205978Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"qX8DQTrChHI1GIGruVTFeR7c8UEpF2b1pwPmJSWe0yh59jZIQHVMahUJAKyhadSPZ2gKXZZn6RpHZiTTn5DbBg==","signature_status":"signed_v1","signed_at":"2026-05-22T02:04:56.206448Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.22817","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:67d014b285ab832f8adff12a98bb1d9a5b2c47acf9d754ff4e0269eeddabd1b7","sha256:fcb0b1ca10bc36cbce43075484de5d2519bf38f298bf6f68bfaf37bef1eeaff4"],"state_sha256":"c23cbc4c7a9f5e320c4149c0d3a9bd6aceeb65782e5c897787fb53eecee619b7"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"G2FGXT4a3kPRKIeNDxekS4i01pu8QQyR6INPdqHejWvNw4L6JUSbPqykWaPkLQXeAfRiyn86lqwZ2qHFRXioAg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-26T11:10:49.394369Z","bundle_sha256":"404ff72b1e117f3d82557f8aca7c8ffada6bf7182329ec7669193d7dbee64a56"}}