{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2023:FWTBR3PM2JJE6N5KHNU2EBUHAG","short_pith_number":"pith:FWTBR3PM","canonical_record":{"source":{"id":"2302.03202","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-02-07T02:24:30Z","cross_cats_sorted":[],"title_canon_sha256":"d8131283e11b380d1dc9ee6f3adfb8b005a0f8126f2a8794ca24fad8520db9fd","abstract_canon_sha256":"8d4beac6ea37495242cf338c2e052feb064805e8d4d7e73e30cabecc84f172fb"},"schema_version":"1.0"},"canonical_sha256":"2da618edecd2524f37aa3b69a2068701bf8bffb3997c81a1ec82c5e313c507ab","source":{"kind":"arxiv","id":"2302.03202","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2302.03202","created_at":"2026-07-05T05:40:11Z"},{"alias_kind":"arxiv_version","alias_value":"2302.03202v2","created_at":"2026-07-05T05:40:11Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2302.03202","created_at":"2026-07-05T05:40:11Z"},{"alias_kind":"pith_short_12","alias_value":"FWTBR3PM2JJE","created_at":"2026-07-05T05:40:11Z"},{"alias_kind":"pith_short_16","alias_value":"FWTBR3PM2JJE6N5K","created_at":"2026-07-05T05:40:11Z"},{"alias_kind":"pith_short_8","alias_value":"FWTBR3PM","created_at":"2026-07-05T05:40:11Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2023:FWTBR3PM2JJE6N5KHNU2EBUHAG","target":"record","payload":{"canonical_record":{"source":{"id":"2302.03202","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-02-07T02:24:30Z","cross_cats_sorted":[],"title_canon_sha256":"d8131283e11b380d1dc9ee6f3adfb8b005a0f8126f2a8794ca24fad8520db9fd","abstract_canon_sha256":"8d4beac6ea37495242cf338c2e052feb064805e8d4d7e73e30cabecc84f172fb"},"schema_version":"1.0"},"canonical_sha256":"2da618edecd2524f37aa3b69a2068701bf8bffb3997c81a1ec82c5e313c507ab","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:40:11.534448Z","signature_b64":"zpPUTT4/69lLQvshnIGgILqjUOiJ6008c+kE3t9TBrYSYkteYe+DCRoRrmsp6P70gHGyp6qEpBgP8uZ3q+KPDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2da618edecd2524f37aa3b69a2068701bf8bffb3997c81a1ec82c5e313c507ab","last_reissued_at":"2026-07-05T05:40:11.533962Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:40:11.533962Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2302.03202","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":"2026-07-05T05:40:11Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ifoidUuBpjZsGNooKRSZNRuVL88wge6PZpLW8QPLWB8gv99MMkUJrPruj7qJxVGfOyRDgfW/zDtDojT1c/ohCA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-22T13:56:08.874691Z"},"content_sha256":"707e3567832444e72df8548c4659d7ef96a014b4cb64129f38c6f2e66a24669e","schema_version":"1.0","event_id":"sha256:707e3567832444e72df8548c4659d7ef96a014b4cb64129f38c6f2e66a24669e"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2023:FWTBR3PM2JJE6N5KHNU2EBUHAG","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Exploring the Benefits of Training Expert Language Models over Instruction Tuning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Doyoung Kim, Joel Jang, Kyungjae Lee, Lajanugen Logeswaran, Minjoon Seo, Moontae Lee, Seonghyeon Ye, Seungone Kim","submitted_at":"2023-02-07T02:24:30Z","abstract_excerpt":"Recently, Language Models (LMs) instruction-tuned on multiple tasks, also known as multitask-prompted fine-tuning (MT), have shown the capability to generalize to unseen tasks. Previous work has shown that scaling the number of training tasks is the key component in making stronger MT LMs. In this work, we report an unexpected finding that an expert LM fine-tuned on just a single task can outperform an MT LM trained with 300+ different tasks on 11 different unseen datasets and on 13 datasets of the BIG-bench benchmark by a mean accuracy of 3.20% and 1.29%, respectively. This finding casts doub"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2302.03202","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/2302.03202/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":"2026-07-05T05:40:11Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"wLZZoVSto2B5mgBsM6vrwod5wEf4UMaRMgeu0WyjALmxuGas07PLXpDpttc5C1Bum4A/tSVBB/ZmJD/TMCyxAA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-22T13:56:08.875200Z"},"content_sha256":"5271104762edc8fa828262417c32f350a8605944f49f3562094465cf0add68c3","schema_version":"1.0","event_id":"sha256:5271104762edc8fa828262417c32f350a8605944f49f3562094465cf0add68c3"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/FWTBR3PM2JJE6N5KHNU2EBUHAG/bundle.json","state_url":"https://pith.science/pith/FWTBR3PM2JJE6N5KHNU2EBUHAG/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/FWTBR3PM2JJE6N5KHNU2EBUHAG/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":"2026-08-22T13:56:08Z","links":{"resolver":"https://pith.science/pith/FWTBR3PM2JJE6N5KHNU2EBUHAG","bundle":"https://pith.science/pith/FWTBR3PM2JJE6N5KHNU2EBUHAG/bundle.json","state":"https://pith.science/pith/FWTBR3PM2JJE6N5KHNU2EBUHAG/state.json","well_known_bundle":"https://pith.science/.well-known/pith/FWTBR3PM2JJE6N5KHNU2EBUHAG/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2023:FWTBR3PM2JJE6N5KHNU2EBUHAG","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":"8d4beac6ea37495242cf338c2e052feb064805e8d4d7e73e30cabecc84f172fb","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-02-07T02:24:30Z","title_canon_sha256":"d8131283e11b380d1dc9ee6f3adfb8b005a0f8126f2a8794ca24fad8520db9fd"},"schema_version":"1.0","source":{"id":"2302.03202","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2302.03202","created_at":"2026-07-05T05:40:11Z"},{"alias_kind":"arxiv_version","alias_value":"2302.03202v2","created_at":"2026-07-05T05:40:11Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2302.03202","created_at":"2026-07-05T05:40:11Z"},{"alias_kind":"pith_short_12","alias_value":"FWTBR3PM2JJE","created_at":"2026-07-05T05:40:11Z"},{"alias_kind":"pith_short_16","alias_value":"FWTBR3PM2JJE6N5K","created_at":"2026-07-05T05:40:11Z"},{"alias_kind":"pith_short_8","alias_value":"FWTBR3PM","created_at":"2026-07-05T05:40:11Z"}],"graph_snapshots":[{"event_id":"sha256:5271104762edc8fa828262417c32f350a8605944f49f3562094465cf0add68c3","target":"graph","created_at":"2026-07-05T05:40:11Z","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/2302.03202/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Recently, Language Models (LMs) instruction-tuned on multiple tasks, also known as multitask-prompted fine-tuning (MT), have shown the capability to generalize to unseen tasks. Previous work has shown that scaling the number of training tasks is the key component in making stronger MT LMs. In this work, we report an unexpected finding that an expert LM fine-tuned on just a single task can outperform an MT LM trained with 300+ different tasks on 11 different unseen datasets and on 13 datasets of the BIG-bench benchmark by a mean accuracy of 3.20% and 1.29%, respectively. This finding casts doub","authors_text":"Doyoung Kim, Joel Jang, Kyungjae Lee, Lajanugen Logeswaran, Minjoon Seo, Moontae Lee, Seonghyeon Ye, Seungone Kim","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-02-07T02:24:30Z","title":"Exploring the Benefits of Training Expert Language Models over Instruction Tuning"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2302.03202","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":"sha256:707e3567832444e72df8548c4659d7ef96a014b4cb64129f38c6f2e66a24669e","target":"record","created_at":"2026-07-05T05:40:11Z","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":"8d4beac6ea37495242cf338c2e052feb064805e8d4d7e73e30cabecc84f172fb","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-02-07T02:24:30Z","title_canon_sha256":"d8131283e11b380d1dc9ee6f3adfb8b005a0f8126f2a8794ca24fad8520db9fd"},"schema_version":"1.0","source":{"id":"2302.03202","kind":"arxiv","version":2}},"canonical_sha256":"2da618edecd2524f37aa3b69a2068701bf8bffb3997c81a1ec82c5e313c507ab","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"2da618edecd2524f37aa3b69a2068701bf8bffb3997c81a1ec82c5e313c507ab","first_computed_at":"2026-07-05T05:40:11.533962Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T05:40:11.533962Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"zpPUTT4/69lLQvshnIGgILqjUOiJ6008c+kE3t9TBrYSYkteYe+DCRoRrmsp6P70gHGyp6qEpBgP8uZ3q+KPDQ==","signature_status":"signed_v1","signed_at":"2026-07-05T05:40:11.534448Z","signed_message":"canonical_sha256_bytes"},"source_id":"2302.03202","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:707e3567832444e72df8548c4659d7ef96a014b4cb64129f38c6f2e66a24669e","sha256:5271104762edc8fa828262417c32f350a8605944f49f3562094465cf0add68c3"],"state_sha256":"e1ac82b36f305c2df6f1740c800ba163385ef91ab6d67d76451ce7dab02f34c6"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"68+4xo20EmoBoyeVY04pOk1EXAYUx330LJTHtmIChCjyp7fuiRwa2pdAHPfAx/FdkaZPdvbFf2YbG8ZpFSmQCQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-22T13:56:08.880051Z","bundle_sha256":"d171e8ea85ea39d605dff1c82ec280be5313a420636649422dbe0de3de60d7fd"}}