{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2023:DK5FBX4YCFJ6G2TWFSB26HL7AP","short_pith_number":"pith:DK5FBX4Y","canonical_record":{"source":{"id":"2311.09476","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-11-16T00:39:39Z","cross_cats_sorted":["cs.AI","cs.IR"],"title_canon_sha256":"c073176ea0a6ef0197234f80d2f652b45562c20de76036962ed925ba2aa4e0f1","abstract_canon_sha256":"1e819053b3af60267520c5efed060e8a7552da60bd876925b37ac2fad31f9d0c"},"schema_version":"1.0"},"canonical_sha256":"1aba50df981153e36a762c83af1d7f03fb6f739d7c1f27ab15769ebac2161653","source":{"kind":"arxiv","id":"2311.09476","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2311.09476","created_at":"2026-07-05T08:02:48Z"},{"alias_kind":"arxiv_version","alias_value":"2311.09476v2","created_at":"2026-07-05T08:02:48Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2311.09476","created_at":"2026-07-05T08:02:48Z"},{"alias_kind":"pith_short_12","alias_value":"DK5FBX4YCFJ6","created_at":"2026-07-05T08:02:48Z"},{"alias_kind":"pith_short_16","alias_value":"DK5FBX4YCFJ6G2TW","created_at":"2026-07-05T08:02:48Z"},{"alias_kind":"pith_short_8","alias_value":"DK5FBX4Y","created_at":"2026-07-05T08:02:48Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2023:DK5FBX4YCFJ6G2TWFSB26HL7AP","target":"record","payload":{"canonical_record":{"source":{"id":"2311.09476","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-11-16T00:39:39Z","cross_cats_sorted":["cs.AI","cs.IR"],"title_canon_sha256":"c073176ea0a6ef0197234f80d2f652b45562c20de76036962ed925ba2aa4e0f1","abstract_canon_sha256":"1e819053b3af60267520c5efed060e8a7552da60bd876925b37ac2fad31f9d0c"},"schema_version":"1.0"},"canonical_sha256":"1aba50df981153e36a762c83af1d7f03fb6f739d7c1f27ab15769ebac2161653","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:02:48.009584Z","signature_b64":"8INLj2u9RijAeS3S8DT3TDAeWD9uUSqWpx/V06pn640qGZJupZKR1s9wM2ecBih2L2huU6oMnRZnn8QEYkG4CQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1aba50df981153e36a762c83af1d7f03fb6f739d7c1f27ab15769ebac2161653","last_reissued_at":"2026-07-05T08:02:48.009104Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:02:48.009104Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2311.09476","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T08:02:48Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"THNM7RL36ZqDoHEZwdW51ydH9nxl+0JcvU8fYskenjaprwsdgJxfdq/Dn9e6oxZW2UBHq6Exbv2MA68Xwkd0Dg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-16T01:08:05.449146Z"},"content_sha256":"4e3416f750b3411f39f418ba7f93fae50a1b7acaf4b7b3dcb2a8ce4b12fddc09","schema_version":"1.0","event_id":"sha256:4e3416f750b3411f39f418ba7f93fae50a1b7acaf4b7b3dcb2a8ce4b12fddc09"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2023:DK5FBX4YCFJ6G2TWFSB26HL7AP","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"ARES: An Automated Evaluation Framework for Retrieval-Augmented Generation Systems","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.IR"],"primary_cat":"cs.CL","authors_text":"Christopher Potts, Jon Saad-Falcon, Matei Zaharia, Omar Khattab","submitted_at":"2023-11-16T00:39:39Z","abstract_excerpt":"Evaluating retrieval-augmented generation (RAG) systems traditionally relies on hand annotations for input queries, passages to retrieve, and responses to generate. We introduce ARES, an Automated RAG Evaluation System, for evaluating RAG systems along the dimensions of context relevance, answer faithfulness, and answer relevance. By creating its own synthetic training data, ARES finetunes lightweight LM judges to assess the quality of individual RAG components. To mitigate potential prediction errors, ARES utilizes a small set of human-annotated datapoints for prediction-powered inference (PP"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2311.09476","kind":"arxiv","version":2},"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/2311.09476/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":"2026-07-05T08:02:48Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"B5r0EqzEL5/WhbUe87Oxh74W0bZbm66MZh3nSMq2twets7EMOt2GZgq6FtmImrmqP/J+otNNvKUjdqC+qTtZBg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-16T01:08:05.449546Z"},"content_sha256":"dba2b7e8046b35e164d0d8e532543233d2b035d40e17c2921622b38d6d50312c","schema_version":"1.0","event_id":"sha256:dba2b7e8046b35e164d0d8e532543233d2b035d40e17c2921622b38d6d50312c"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/DK5FBX4YCFJ6G2TWFSB26HL7AP/bundle.json","state_url":"https://pith.science/pith/DK5FBX4YCFJ6G2TWFSB26HL7AP/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/DK5FBX4YCFJ6G2TWFSB26HL7AP/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":"2026-07-16T01:08:05Z","links":{"resolver":"https://pith.science/pith/DK5FBX4YCFJ6G2TWFSB26HL7AP","bundle":"https://pith.science/pith/DK5FBX4YCFJ6G2TWFSB26HL7AP/bundle.json","state":"https://pith.science/pith/DK5FBX4YCFJ6G2TWFSB26HL7AP/state.json","well_known_bundle":"https://pith.science/.well-known/pith/DK5FBX4YCFJ6G2TWFSB26HL7AP/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2023:DK5FBX4YCFJ6G2TWFSB26HL7AP","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":"1e819053b3af60267520c5efed060e8a7552da60bd876925b37ac2fad31f9d0c","cross_cats_sorted":["cs.AI","cs.IR"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-11-16T00:39:39Z","title_canon_sha256":"c073176ea0a6ef0197234f80d2f652b45562c20de76036962ed925ba2aa4e0f1"},"schema_version":"1.0","source":{"id":"2311.09476","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2311.09476","created_at":"2026-07-05T08:02:48Z"},{"alias_kind":"arxiv_version","alias_value":"2311.09476v2","created_at":"2026-07-05T08:02:48Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2311.09476","created_at":"2026-07-05T08:02:48Z"},{"alias_kind":"pith_short_12","alias_value":"DK5FBX4YCFJ6","created_at":"2026-07-05T08:02:48Z"},{"alias_kind":"pith_short_16","alias_value":"DK5FBX4YCFJ6G2TW","created_at":"2026-07-05T08:02:48Z"},{"alias_kind":"pith_short_8","alias_value":"DK5FBX4Y","created_at":"2026-07-05T08:02:48Z"}],"graph_snapshots":[{"event_id":"sha256:dba2b7e8046b35e164d0d8e532543233d2b035d40e17c2921622b38d6d50312c","target":"graph","created_at":"2026-07-05T08:02:48Z","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/2311.09476/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Evaluating retrieval-augmented generation (RAG) systems traditionally relies on hand annotations for input queries, passages to retrieve, and responses to generate. We introduce ARES, an Automated RAG Evaluation System, for evaluating RAG systems along the dimensions of context relevance, answer faithfulness, and answer relevance. By creating its own synthetic training data, ARES finetunes lightweight LM judges to assess the quality of individual RAG components. To mitigate potential prediction errors, ARES utilizes a small set of human-annotated datapoints for prediction-powered inference (PP","authors_text":"Christopher Potts, Jon Saad-Falcon, Matei Zaharia, Omar Khattab","cross_cats":["cs.AI","cs.IR"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-11-16T00:39:39Z","title":"ARES: An Automated Evaluation Framework for Retrieval-Augmented Generation Systems"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2311.09476","kind":"arxiv","version":2},"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":"sha256:4e3416f750b3411f39f418ba7f93fae50a1b7acaf4b7b3dcb2a8ce4b12fddc09","target":"record","created_at":"2026-07-05T08:02:48Z","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":"1e819053b3af60267520c5efed060e8a7552da60bd876925b37ac2fad31f9d0c","cross_cats_sorted":["cs.AI","cs.IR"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-11-16T00:39:39Z","title_canon_sha256":"c073176ea0a6ef0197234f80d2f652b45562c20de76036962ed925ba2aa4e0f1"},"schema_version":"1.0","source":{"id":"2311.09476","kind":"arxiv","version":2}},"canonical_sha256":"1aba50df981153e36a762c83af1d7f03fb6f739d7c1f27ab15769ebac2161653","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"1aba50df981153e36a762c83af1d7f03fb6f739d7c1f27ab15769ebac2161653","first_computed_at":"2026-07-05T08:02:48.009104Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:02:48.009104Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"8INLj2u9RijAeS3S8DT3TDAeWD9uUSqWpx/V06pn640qGZJupZKR1s9wM2ecBih2L2huU6oMnRZnn8QEYkG4CQ==","signature_status":"signed_v1","signed_at":"2026-07-05T08:02:48.009584Z","signed_message":"canonical_sha256_bytes"},"source_id":"2311.09476","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:4e3416f750b3411f39f418ba7f93fae50a1b7acaf4b7b3dcb2a8ce4b12fddc09","sha256:dba2b7e8046b35e164d0d8e532543233d2b035d40e17c2921622b38d6d50312c"],"state_sha256":"23bd3e7cd251b4cd99eee526e0d0ddd7bef55637eda892686aa223887f3c2c78"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"BqzUHLBZIwVxrj/SwvRbAdnw2ZNUeicOkwpAU23NhwCCjunAiWCl4zXobjfjGIurB2zohcFogaF5WFJ8v2puAQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-16T01:08:05.452142Z","bundle_sha256":"9bd34965ffb86aa52980b47892752d36a111957622c51d407f93bd10a50509fc"}}