{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:AXO2O2ENIVDJS72VAV745DMXNI","short_pith_number":"pith:AXO2O2EN","canonical_record":{"source":{"id":"2404.13940","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-04-22T07:32:03Z","cross_cats_sorted":[],"title_canon_sha256":"1a37feb76d4fb36443a27a133c7b58f05505cb71b258322af61147e5e0e4ca6f","abstract_canon_sha256":"78b271a93fcc265c234a6b8d51a4659a3052defa885e54a9a39d8a3a26e6bb67"},"schema_version":"1.0"},"canonical_sha256":"05dda7688d4546997f55057fce8d976a1499c74afc626c7527174fcd53da0e82","source":{"kind":"arxiv","id":"2404.13940","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2404.13940","created_at":"2026-07-05T09:09:21Z"},{"alias_kind":"arxiv_version","alias_value":"2404.13940v3","created_at":"2026-07-05T09:09:21Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2404.13940","created_at":"2026-07-05T09:09:21Z"},{"alias_kind":"pith_short_12","alias_value":"AXO2O2ENIVDJ","created_at":"2026-07-05T09:09:21Z"},{"alias_kind":"pith_short_16","alias_value":"AXO2O2ENIVDJS72V","created_at":"2026-07-05T09:09:21Z"},{"alias_kind":"pith_short_8","alias_value":"AXO2O2EN","created_at":"2026-07-05T09:09:21Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:AXO2O2ENIVDJS72VAV745DMXNI","target":"record","payload":{"canonical_record":{"source":{"id":"2404.13940","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-04-22T07:32:03Z","cross_cats_sorted":[],"title_canon_sha256":"1a37feb76d4fb36443a27a133c7b58f05505cb71b258322af61147e5e0e4ca6f","abstract_canon_sha256":"78b271a93fcc265c234a6b8d51a4659a3052defa885e54a9a39d8a3a26e6bb67"},"schema_version":"1.0"},"canonical_sha256":"05dda7688d4546997f55057fce8d976a1499c74afc626c7527174fcd53da0e82","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:09:21.867462Z","signature_b64":"x7ITjsJ6AP+KzztiX9dMWo4jgzOmFocygRw3nv36Epayqbags7wWvICLcYXWzuHuBpRubZEw4XTXtrb+RaRuAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"05dda7688d4546997f55057fce8d976a1499c74afc626c7527174fcd53da0e82","last_reissued_at":"2026-07-05T09:09:21.866940Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:09:21.866940Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2404.13940","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T09:09:21Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ESyoOePr31DznPRC+2ZHrllTg+xZu2Z7uRylWKLlIsYX6BSG9pfI75zPXAgET4mbGFNBFvM9Nbow0UKK9kkQAw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-11T20:47:56.269150Z"},"content_sha256":"96c2482c0402e0caa3853f39f87d49e1188e7812ac9275a74e95af06d1407bca","schema_version":"1.0","event_id":"sha256:96c2482c0402e0caa3853f39f87d49e1188e7812ac9275a74e95af06d1407bca"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:AXO2O2ENIVDJS72VAV745DMXNI","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"A User-Centric Multi-Intent Benchmark for Evaluating Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Fengran Mo, Jian-Yun Nie, Jiayin Wang, Min Zhang, Peijie Sun, Weizhi Ma","submitted_at":"2024-04-22T07:32:03Z","abstract_excerpt":"Large language models (LLMs) are essential tools that users employ across various scenarios, so evaluating their performance and guiding users in selecting the suitable service is important. Although many benchmarks exist, they mainly focus on specific predefined model abilities, such as world knowledge, reasoning, etc. Based on these ability scores, it is hard for users to determine which LLM best suits their particular needs. To address these issues, we propose to evaluate LLMs from a user-centric perspective and design this benchmark to measure their efficacy in satisfying user needs under "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.13940","kind":"arxiv","version":3},"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/2404.13940/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":"2026-07-05T09:09:21Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"/pMHErWpk9zR01gqTNJnrjDk/N6HUBqO6SNOR643Wa25OdQ2ro5qD+o/IXXAR+tchwWvMEh4CEbavgQBE2D1DA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-11T20:47:56.269645Z"},"content_sha256":"3472acd14a228c50a435fb44cfdfb0fb4f03d521fefcac714c3331cd80c823d1","schema_version":"1.0","event_id":"sha256:3472acd14a228c50a435fb44cfdfb0fb4f03d521fefcac714c3331cd80c823d1"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/AXO2O2ENIVDJS72VAV745DMXNI/bundle.json","state_url":"https://pith.science/pith/AXO2O2ENIVDJS72VAV745DMXNI/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/AXO2O2ENIVDJS72VAV745DMXNI/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":"2026-08-11T20:47:56Z","links":{"resolver":"https://pith.science/pith/AXO2O2ENIVDJS72VAV745DMXNI","bundle":"https://pith.science/pith/AXO2O2ENIVDJS72VAV745DMXNI/bundle.json","state":"https://pith.science/pith/AXO2O2ENIVDJS72VAV745DMXNI/state.json","well_known_bundle":"https://pith.science/.well-known/pith/AXO2O2ENIVDJS72VAV745DMXNI/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:AXO2O2ENIVDJS72VAV745DMXNI","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":"78b271a93fcc265c234a6b8d51a4659a3052defa885e54a9a39d8a3a26e6bb67","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-04-22T07:32:03Z","title_canon_sha256":"1a37feb76d4fb36443a27a133c7b58f05505cb71b258322af61147e5e0e4ca6f"},"schema_version":"1.0","source":{"id":"2404.13940","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2404.13940","created_at":"2026-07-05T09:09:21Z"},{"alias_kind":"arxiv_version","alias_value":"2404.13940v3","created_at":"2026-07-05T09:09:21Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2404.13940","created_at":"2026-07-05T09:09:21Z"},{"alias_kind":"pith_short_12","alias_value":"AXO2O2ENIVDJ","created_at":"2026-07-05T09:09:21Z"},{"alias_kind":"pith_short_16","alias_value":"AXO2O2ENIVDJS72V","created_at":"2026-07-05T09:09:21Z"},{"alias_kind":"pith_short_8","alias_value":"AXO2O2EN","created_at":"2026-07-05T09:09:21Z"}],"graph_snapshots":[{"event_id":"sha256:3472acd14a228c50a435fb44cfdfb0fb4f03d521fefcac714c3331cd80c823d1","target":"graph","created_at":"2026-07-05T09:09:21Z","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/2404.13940/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Large language models (LLMs) are essential tools that users employ across various scenarios, so evaluating their performance and guiding users in selecting the suitable service is important. Although many benchmarks exist, they mainly focus on specific predefined model abilities, such as world knowledge, reasoning, etc. Based on these ability scores, it is hard for users to determine which LLM best suits their particular needs. To address these issues, we propose to evaluate LLMs from a user-centric perspective and design this benchmark to measure their efficacy in satisfying user needs under ","authors_text":"Fengran Mo, Jian-Yun Nie, Jiayin Wang, Min Zhang, Peijie Sun, Weizhi Ma","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-04-22T07:32:03Z","title":"A User-Centric Multi-Intent Benchmark for Evaluating Large Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.13940","kind":"arxiv","version":3},"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":"sha256:96c2482c0402e0caa3853f39f87d49e1188e7812ac9275a74e95af06d1407bca","target":"record","created_at":"2026-07-05T09:09:21Z","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":"78b271a93fcc265c234a6b8d51a4659a3052defa885e54a9a39d8a3a26e6bb67","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-04-22T07:32:03Z","title_canon_sha256":"1a37feb76d4fb36443a27a133c7b58f05505cb71b258322af61147e5e0e4ca6f"},"schema_version":"1.0","source":{"id":"2404.13940","kind":"arxiv","version":3}},"canonical_sha256":"05dda7688d4546997f55057fce8d976a1499c74afc626c7527174fcd53da0e82","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"05dda7688d4546997f55057fce8d976a1499c74afc626c7527174fcd53da0e82","first_computed_at":"2026-07-05T09:09:21.866940Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:09:21.866940Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"x7ITjsJ6AP+KzztiX9dMWo4jgzOmFocygRw3nv36Epayqbags7wWvICLcYXWzuHuBpRubZEw4XTXtrb+RaRuAw==","signature_status":"signed_v1","signed_at":"2026-07-05T09:09:21.867462Z","signed_message":"canonical_sha256_bytes"},"source_id":"2404.13940","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:96c2482c0402e0caa3853f39f87d49e1188e7812ac9275a74e95af06d1407bca","sha256:3472acd14a228c50a435fb44cfdfb0fb4f03d521fefcac714c3331cd80c823d1"],"state_sha256":"57eb85a20d7d1819e9ce4f6bc183901368bc62aa545373c023ffdbd7041621da"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"V69A6hPyvvd3qy8XxHL+PKNfGDBWXyI8Tr/014//veWNmtmrOhfT9/8Eowgp5w7HJ1FGTjVWJfpi89X2ZutMBQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-11T20:47:56.283907Z","bundle_sha256":"025e244c94298fa48fc9db0404872647f9f794866ad0d5ebcf99de77bdf8f200"}}