{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:YOK4N5EJZFEF4NBH3Y7BKJ7GYV","short_pith_number":"pith:YOK4N5EJ","schema_version":"1.0","canonical_sha256":"c395c6f489c9485e3427de3e1527e6c576250939a408b583546e20b334776b7f","source":{"kind":"arxiv","id":"2501.17195","version":1},"attestation_state":"computed","paper":{"title":"Atla Selene Mini: A General Purpose Evaluation Model","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Andrei Alexandru, Antonia Calvi, Henry Broomfield, Jackson Golden, Kyle Dai, Mathias Leys, Maurice Burger, Max Bartolo, Roman Engeler, Sashank Pisupati, Toby Drane, Young Sun Park","submitted_at":"2025-01-27T15:09:08Z","abstract_excerpt":"We introduce Atla Selene Mini, a state-of-the-art small language model-as-a-judge (SLMJ). Selene Mini is a general-purpose evaluator that outperforms the best SLMJs and GPT-4o-mini on overall performance across 11 out-of-distribution benchmarks, spanning absolute scoring, classification, and pairwise preference tasks. It is the highest-scoring 8B generative model on RewardBench, surpassing strong baselines like GPT-4o and specialized judges. To achieve this, we develop a principled data curation strategy that augments public datasets with synthetically generated critiques and ensures high qual"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2501.17195","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-01-27T15:09:08Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"9ec14e8d139f2c00f61b97d6d097c12c6a734053fa485bd44ebc1dccd9114922","abstract_canon_sha256":"16878597879e2f0ae3f3f9846e791c2219399bca9b92899ca2531b1f254d7094"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:06:48.965033Z","signature_b64":"8bal6DJZd5A+NRRP0AAUESSh8Iqb92JfVV4BQDoROZKYqFI7jmGMbVXRiYkpxMvFHRAQovUHriFcc0z24ydoBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c395c6f489c9485e3427de3e1527e6c576250939a408b583546e20b334776b7f","last_reissued_at":"2026-07-05T10:06:48.964555Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:06:48.964555Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Atla Selene Mini: A General Purpose Evaluation Model","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Andrei Alexandru, Antonia Calvi, Henry Broomfield, Jackson Golden, Kyle Dai, Mathias Leys, Maurice Burger, Max Bartolo, Roman Engeler, Sashank Pisupati, Toby Drane, Young Sun Park","submitted_at":"2025-01-27T15:09:08Z","abstract_excerpt":"We introduce Atla Selene Mini, a state-of-the-art small language model-as-a-judge (SLMJ). Selene Mini is a general-purpose evaluator that outperforms the best SLMJs and GPT-4o-mini on overall performance across 11 out-of-distribution benchmarks, spanning absolute scoring, classification, and pairwise preference tasks. It is the highest-scoring 8B generative model on RewardBench, surpassing strong baselines like GPT-4o and specialized judges. To achieve this, we develop a principled data curation strategy that augments public datasets with synthetically generated critiques and ensures high qual"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.17195","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/2501.17195/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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2501.17195","created_at":"2026-07-05T10:06:48.964613+00:00"},{"alias_kind":"arxiv_version","alias_value":"2501.17195v1","created_at":"2026-07-05T10:06:48.964613+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.17195","created_at":"2026-07-05T10:06:48.964613+00:00"},{"alias_kind":"pith_short_12","alias_value":"YOK4N5EJZFEF","created_at":"2026-07-05T10:06:48.964613+00:00"},{"alias_kind":"pith_short_16","alias_value":"YOK4N5EJZFEF4NBH","created_at":"2026-07-05T10:06:48.964613+00:00"},{"alias_kind":"pith_short_8","alias_value":"YOK4N5EJ","created_at":"2026-07-05T10:06:48.964613+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":6,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.21627","citing_title":"Counsel: A Meta-Evaluation Dataset for Agentic Tasks","ref_index":23,"is_internal_anchor":false},{"citing_arxiv_id":"2606.03128","citing_title":"Decoupled Smart Contract Audits: Lightweight LLM Framework via Distillation and Aggregation","ref_index":44,"is_internal_anchor":false},{"citing_arxiv_id":"2606.01034","citing_title":"A Finite-Calibration Regime Map for LLM Judge Panels","ref_index":2,"is_internal_anchor":false},{"citing_arxiv_id":"2510.24235","citing_title":"PaTaRM: Bridging Pairwise and Pointwise Signals via Preference-Aware Task-Adaptive Reward Modeling","ref_index":1,"is_internal_anchor":false},{"citing_arxiv_id":"2605.11334","citing_title":"VERDI: Single-Call Confidence Estimation for Verification-Based LLM Judges via Decomposed Inference","ref_index":1,"is_internal_anchor":false},{"citing_arxiv_id":"2604.12312","citing_title":"CompliBench: Benchmarking LLM Judges for Compliance Violation Detection in Dialogue Systems","ref_index":1,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/YOK4N5EJZFEF4NBH3Y7BKJ7GYV","json":"https://pith.science/pith/YOK4N5EJZFEF4NBH3Y7BKJ7GYV.json","graph_json":"https://pith.science/api/pith-number/YOK4N5EJZFEF4NBH3Y7BKJ7GYV/graph.json","events_json":"https://pith.science/api/pith-number/YOK4N5EJZFEF4NBH3Y7BKJ7GYV/events.json","paper":"https://pith.science/paper/YOK4N5EJ"},"agent_actions":{"view_html":"https://pith.science/pith/YOK4N5EJZFEF4NBH3Y7BKJ7GYV","download_json":"https://pith.science/pith/YOK4N5EJZFEF4NBH3Y7BKJ7GYV.json","view_paper":"https://pith.science/paper/YOK4N5EJ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2501.17195&json=true","fetch_graph":"https://pith.science/api/pith-number/YOK4N5EJZFEF4NBH3Y7BKJ7GYV/graph.json","fetch_events":"https://pith.science/api/pith-number/YOK4N5EJZFEF4NBH3Y7BKJ7GYV/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/YOK4N5EJZFEF4NBH3Y7BKJ7GYV/action/timestamp_anchor","attest_storage":"https://pith.science/pith/YOK4N5EJZFEF4NBH3Y7BKJ7GYV/action/storage_attestation","attest_author":"https://pith.science/pith/YOK4N5EJZFEF4NBH3Y7BKJ7GYV/action/author_attestation","sign_citation":"https://pith.science/pith/YOK4N5EJZFEF4NBH3Y7BKJ7GYV/action/citation_signature","submit_replication":"https://pith.science/pith/YOK4N5EJZFEF4NBH3Y7BKJ7GYV/action/replication_record"}},"created_at":"2026-07-05T10:06:48.964613+00:00","updated_at":"2026-07-05T10:06:48.964613+00:00"}