{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:YYVWELS6HVBXGAAWRSB7XT727X","short_pith_number":"pith:YYVWELS6","schema_version":"1.0","canonical_sha256":"c62b622e5e3d437300168c83fbcffafddf3abb3b640d805f93ee73685101c0a2","source":{"kind":"arxiv","id":"2310.05915","version":1},"attestation_state":"computed","paper":{"title":"FireAct: Toward Language Agent Fine-tuning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Baian Chen, Chang Shu, Ehsan Shareghi, Karthik Narasimhan, Nigel Collier, Shunyu Yao","submitted_at":"2023-10-09T17:58:38Z","abstract_excerpt":"Recent efforts have augmented language models (LMs) with external tools or environments, leading to the development of language agents that can reason and act. However, most of these agents rely on few-shot prompting techniques with off-the-shelf LMs. In this paper, we investigate and argue for the overlooked direction of fine-tuning LMs to obtain language agents. Using a setup of question answering (QA) with a Google search API, we explore a variety of base LMs, prompting methods, fine-tuning data, and QA tasks, and find language agents are consistently improved after fine-tuning their backbo"},"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":"2310.05915","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-10-09T17:58:38Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"8935554831002290c5c5cfc69918c68bf71cfcf5f4b7ce5b55abd81c69139963","abstract_canon_sha256":"84ce32a7eb1b43e2df8cbbbab9c127d80b7328e3f952daa433b48ad83d807156"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:58:51.268782Z","signature_b64":"Y5ZBZ9utkshC7frCZriCMLADQ5VhW9OT3pn8aF3Wy+PUJkwU8ixG10/wmXIP7FQn4KdzxSXxefdSo9uEiV+4AQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c62b622e5e3d437300168c83fbcffafddf3abb3b640d805f93ee73685101c0a2","last_reissued_at":"2026-07-05T06:58:51.268327Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:58:51.268327Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"FireAct: Toward Language Agent Fine-tuning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Baian Chen, Chang Shu, Ehsan Shareghi, Karthik Narasimhan, Nigel Collier, Shunyu Yao","submitted_at":"2023-10-09T17:58:38Z","abstract_excerpt":"Recent efforts have augmented language models (LMs) with external tools or environments, leading to the development of language agents that can reason and act. However, most of these agents rely on few-shot prompting techniques with off-the-shelf LMs. In this paper, we investigate and argue for the overlooked direction of fine-tuning LMs to obtain language agents. Using a setup of question answering (QA) with a Google search API, we explore a variety of base LMs, prompting methods, fine-tuning data, and QA tasks, and find language agents are consistently improved after fine-tuning their backbo"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2310.05915","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/2310.05915/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":"2310.05915","created_at":"2026-07-05T06:58:51.268394+00:00"},{"alias_kind":"arxiv_version","alias_value":"2310.05915v1","created_at":"2026-07-05T06:58:51.268394+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2310.05915","created_at":"2026-07-05T06:58:51.268394+00:00"},{"alias_kind":"pith_short_12","alias_value":"YYVWELS6HVBX","created_at":"2026-07-05T06:58:51.268394+00:00"},{"alias_kind":"pith_short_16","alias_value":"YYVWELS6HVBXGAAW","created_at":"2026-07-05T06:58:51.268394+00:00"},{"alias_kind":"pith_short_8","alias_value":"YYVWELS6","created_at":"2026-07-05T06:58:51.268394+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":35,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.27147","citing_title":"Safe Autoregressive Image Generation with Iterative Self-Improving Codebooks","ref_index":5,"is_internal_anchor":false},{"citing_arxiv_id":"2606.21740","citing_title":"Training the Orchestrator: A Supervised Approach to End-to-End PDDL Planning with LLM Agents","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2607.01942","citing_title":"Atomic Task Graph: A Unified Framework for Agentic Planning and Execution","ref_index":41,"is_internal_anchor":false},{"citing_arxiv_id":"2606.07367","citing_title":"Self-evolving LLM agents with in-distribution Optimization","ref_index":31,"is_internal_anchor":false},{"citing_arxiv_id":"2606.04120","citing_title":"SaliMory: Orchestrating Cognitive Memory for Conversational Agents","ref_index":2,"is_internal_anchor":false},{"citing_arxiv_id":"2606.02372","citing_title":"COMAP: Co-Evolving World Models and Agent Policies for LLM Agents","ref_index":42,"is_internal_anchor":false},{"citing_arxiv_id":"2606.24893","citing_title":"AgentOdyssey: Open-Ended Long-Horizon Text Game Generation for Test-Time Continual Learning Agents","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2606.31229","citing_title":"Agentic-Ideation: Sample Efficient Agentic Trajectories Synthesis for Scientific Ideation Agents","ref_index":1,"is_internal_anchor":false},{"citing_arxiv_id":"2605.10325","citing_title":"Verifiable Process Rewards for Agentic Reasoning","ref_index":1,"is_internal_anchor":false},{"citing_arxiv_id":"2605.24828","citing_title":"Test-Time Deep Thinking to Explore Implicit Rules","ref_index":2,"is_internal_anchor":false},{"citing_arxiv_id":"2605.28774","citing_title":"Agent Explorative Policy Optimization for Multimodal Agentic Reasoning","ref_index":75,"is_internal_anchor":false},{"citing_arxiv_id":"2606.12674","citing_title":"Evoflux: Inference-Time Evolution of Executable Tool Workflows for Compact Agents","ref_index":24,"is_internal_anchor":false},{"citing_arxiv_id":"2502.00955","citing_title":"Efficient Multi-Agent System Training with Data Influence-Oriented Tree Search","ref_index":3,"is_internal_anchor":false},{"citing_arxiv_id":"2605.22502","citing_title":"Compiling Agentic Workflows into LLM Weights: Near-Frontier Quality at Two Orders of Magnitude Less Cost","ref_index":31,"is_internal_anchor":false},{"citing_arxiv_id":"2605.14133","citing_title":"ClawForge: Generating Executable Interactive Benchmarks for Command-Line Agents","ref_index":87,"is_internal_anchor":false},{"citing_arxiv_id":"2605.17352","citing_title":"AMATA: Adaptive Multi-Agent Trajectory Alignment for Knowledge-Intensive Question Answering","ref_index":1,"is_internal_anchor":false},{"citing_arxiv_id":"2605.15207","citing_title":"TeamTR: Trust-Region Fine-Tuning for Multi-Agent LLM Coordination","ref_index":41,"is_internal_anchor":false},{"citing_arxiv_id":"2604.27859","citing_title":"Rethinking Agentic Reinforcement Learning In Large Language Models","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2506.03610","citing_title":"Orak: A Foundational Benchmark for Training and Evaluating LLM Agents on Diverse Video Games","ref_index":42,"is_internal_anchor":false},{"citing_arxiv_id":"2509.02547","citing_title":"The Landscape of Agentic Reinforcement Learning for LLMs: A Survey","ref_index":93,"is_internal_anchor":false},{"citing_arxiv_id":"2511.11362","citing_title":"On-Device Fine-Tuning via Backprop-Free Zeroth-Order Optimization","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2507.02592","citing_title":"WebSailor: Navigating Super-human Reasoning for Web Agent","ref_index":2,"is_internal_anchor":false},{"citing_arxiv_id":"2401.05459","citing_title":"Personal LLM Agents: Insights and Survey about the Capability, Efficiency and Security","ref_index":247,"is_internal_anchor":false},{"citing_arxiv_id":"2605.14133","citing_title":"ClawForge: Generating Executable Interactive Benchmarks for Command-Line Agents","ref_index":87,"is_internal_anchor":false},{"citing_arxiv_id":"2602.20867","citing_title":"SoK: Agentic Skills -- Beyond Tool Use in LLM Agents","ref_index":43,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/YYVWELS6HVBXGAAWRSB7XT727X","json":"https://pith.science/pith/YYVWELS6HVBXGAAWRSB7XT727X.json","graph_json":"https://pith.science/api/pith-number/YYVWELS6HVBXGAAWRSB7XT727X/graph.json","events_json":"https://pith.science/api/pith-number/YYVWELS6HVBXGAAWRSB7XT727X/events.json","paper":"https://pith.science/paper/YYVWELS6"},"agent_actions":{"view_html":"https://pith.science/pith/YYVWELS6HVBXGAAWRSB7XT727X","download_json":"https://pith.science/pith/YYVWELS6HVBXGAAWRSB7XT727X.json","view_paper":"https://pith.science/paper/YYVWELS6","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2310.05915&json=true","fetch_graph":"https://pith.science/api/pith-number/YYVWELS6HVBXGAAWRSB7XT727X/graph.json","fetch_events":"https://pith.science/api/pith-number/YYVWELS6HVBXGAAWRSB7XT727X/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/YYVWELS6HVBXGAAWRSB7XT727X/action/timestamp_anchor","attest_storage":"https://pith.science/pith/YYVWELS6HVBXGAAWRSB7XT727X/action/storage_attestation","attest_author":"https://pith.science/pith/YYVWELS6HVBXGAAWRSB7XT727X/action/author_attestation","sign_citation":"https://pith.science/pith/YYVWELS6HVBXGAAWRSB7XT727X/action/citation_signature","submit_replication":"https://pith.science/pith/YYVWELS6HVBXGAAWRSB7XT727X/action/replication_record"}},"created_at":"2026-07-05T06:58:51.268394+00:00","updated_at":"2026-07-05T06:58:51.268394+00:00"}