{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:7G2KCMHD77CZB2RHZUGJKHT5MX","short_pith_number":"pith:7G2KCMHD","schema_version":"1.0","canonical_sha256":"f9b4a130e3ffc590ea27cd0c951e7d65d54e8492524cf2afed3818d9bb63b9a7","source":{"kind":"arxiv","id":"2607.28657","version":1},"attestation_state":"computed","paper":{"title":"TAPR: Enhancing LLM Performance with a Task-Aware Prompt Rewriter","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Emanuele Bastianelli, Hosein Azarbonyad, Oliver Savolainen","submitted_at":"2026-07-17T15:37:07Z","abstract_excerpt":"Large Language Models (LLMs) often require carefully crafted prompts to unlock their full potential, which can be a barrier for non-expert users. This work addresses the challenge by introducing a Task-Aware Prompt Rewriter (TAPR), a model that reformulates user prompts into task-optimized prompts with the explicit goal of improving downstream LLM performance. We train TAPR using reinforcement learning with Group Relative Policy Optimization (GRPO), where rewards are derived from LLM-as-judge evaluations of both the reformulated prompt and the corresponding task output. Experimental results on"},"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":"2607.28657","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-07-17T15:37:07Z","cross_cats_sorted":[],"title_canon_sha256":"176cac56cc332db21f505cc51bd2f5a545a9aa6be15b604e0f526f230fb53e9d","abstract_canon_sha256":"4f97c0c218e47109867f93d4c6e99efa89c42c9f2a361ff906e93e00d88e8c49"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-03T00:11:55.904673Z","signature_b64":"5UGyfn633EzDeQ7loaadjBdv6fK0xn17L34AmOan0ULrt4VydeyYogSCx2TysQGdnlcpUlmt6MRVfrti79yqDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f9b4a130e3ffc590ea27cd0c951e7d65d54e8492524cf2afed3818d9bb63b9a7","last_reissued_at":"2026-08-03T00:11:55.902294Z","signature_status":"signed_v1","first_computed_at":"2026-08-03T00:11:55.902294Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"TAPR: Enhancing LLM Performance with a Task-Aware Prompt Rewriter","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Emanuele Bastianelli, Hosein Azarbonyad, Oliver Savolainen","submitted_at":"2026-07-17T15:37:07Z","abstract_excerpt":"Large Language Models (LLMs) often require carefully crafted prompts to unlock their full potential, which can be a barrier for non-expert users. This work addresses the challenge by introducing a Task-Aware Prompt Rewriter (TAPR), a model that reformulates user prompts into task-optimized prompts with the explicit goal of improving downstream LLM performance. We train TAPR using reinforcement learning with Group Relative Policy Optimization (GRPO), where rewards are derived from LLM-as-judge evaluations of both the reformulated prompt and the corresponding task output. Experimental results on"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.28657","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/2607.28657/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":"2607.28657","created_at":"2026-08-03T00:11:55.903909+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.28657v1","created_at":"2026-08-03T00:11:55.903909+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.28657","created_at":"2026-08-03T00:11:55.903909+00:00"},{"alias_kind":"pith_short_12","alias_value":"7G2KCMHD77CZ","created_at":"2026-08-03T00:11:55.903909+00:00"},{"alias_kind":"pith_short_16","alias_value":"7G2KCMHD77CZB2RH","created_at":"2026-08-03T00:11:55.903909+00:00"},{"alias_kind":"pith_short_8","alias_value":"7G2KCMHD","created_at":"2026-08-03T00:11:55.903909+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/7G2KCMHD77CZB2RHZUGJKHT5MX","json":"https://pith.science/pith/7G2KCMHD77CZB2RHZUGJKHT5MX.json","graph_json":"https://pith.science/api/pith-number/7G2KCMHD77CZB2RHZUGJKHT5MX/graph.json","events_json":"https://pith.science/api/pith-number/7G2KCMHD77CZB2RHZUGJKHT5MX/events.json","paper":"https://pith.science/paper/7G2KCMHD"},"agent_actions":{"view_html":"https://pith.science/pith/7G2KCMHD77CZB2RHZUGJKHT5MX","download_json":"https://pith.science/pith/7G2KCMHD77CZB2RHZUGJKHT5MX.json","view_paper":"https://pith.science/paper/7G2KCMHD","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.28657&json=true","fetch_graph":"https://pith.science/api/pith-number/7G2KCMHD77CZB2RHZUGJKHT5MX/graph.json","fetch_events":"https://pith.science/api/pith-number/7G2KCMHD77CZB2RHZUGJKHT5MX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/7G2KCMHD77CZB2RHZUGJKHT5MX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/7G2KCMHD77CZB2RHZUGJKHT5MX/action/storage_attestation","attest_author":"https://pith.science/pith/7G2KCMHD77CZB2RHZUGJKHT5MX/action/author_attestation","sign_citation":"https://pith.science/pith/7G2KCMHD77CZB2RHZUGJKHT5MX/action/citation_signature","submit_replication":"https://pith.science/pith/7G2KCMHD77CZB2RHZUGJKHT5MX/action/replication_record"}},"created_at":"2026-08-03T00:11:55.903909+00:00","updated_at":"2026-08-03T00:11:55.903909+00:00"}