{"work":{"id":"778f739e-5f55-4961-8a2a-e4736a2757f4","openalex_id":null,"doi":null,"arxiv_id":"2303.11366","raw_key":null,"title":"Reflexion: Language Agents with Verbal Reinforcement Learning","authors":null,"authors_text":"Noah Shinn, Federico Cassano, Edward Berman, Ashwin Gopinath, Karthik Narasimhan, Shunyu Yao","year":2023,"venue":"cs.AI","abstract":"Large language models (LLMs) have been increasingly used to interact with external environments (e.g., games, compilers, APIs) as goal-driven agents. However, it remains challenging for these language agents to quickly and efficiently learn from trial-and-error as traditional reinforcement learning methods require extensive training samples and expensive model fine-tuning. We propose Reflexion, a novel framework to reinforce language agents not by updating weights, but instead through linguistic feedback. Concretely, Reflexion agents verbally reflect on task feedback signals, then maintain their own reflective text in an episodic memory buffer to induce better decision-making in subsequent trials. Reflexion is flexible enough to incorporate various types (scalar values or free-form language) and sources (external or internally simulated) of feedback signals, and obtains significant improvements over a baseline agent across diverse tasks (sequential decision-making, coding, language reasoning). For example, Reflexion achieves a 91% pass@1 accuracy on the HumanEval coding benchmark, surpassing the previous state-of-the-art GPT-4 that achieves 80%. We also conduct ablation and analysis studies using different feedback signals, feedback incorporation methods, and agent types, and provide insights into how they affect performance.","external_url":"https://arxiv.org/abs/2303.11366","cited_by_count":null,"metadata_source":"pith","metadata_fetched_at":"2026-05-14T22:18:03.990313+00:00","pith_arxiv_id":"2303.11366","created_at":"2026-05-09T05:50:26.094618+00:00","updated_at":"2026-05-14T22:18:03.990313+00:00","title_quality_ok":true,"display_title":"Reflexion: Language Agents with Verbal Reinforcement Learning","render_title":"Reflexion: Language Agents with Verbal Reinforcement Learning"},"hub":{"state":{"work_id":"778f739e-5f55-4961-8a2a-e4736a2757f4","tier":"hub","tier_reason":"10+ Pith inbound or 1,000+ external citations","pith_inbound_count":81,"external_cited_by_count":null,"distinct_field_count":13,"first_pith_cited_at":"2023-03-30T18:30:01+00:00","last_pith_cited_at":"2026-05-12T17:58:20+00:00","author_build_status":"not_needed","summary_status":"needed","contexts_status":"needed","graph_status":"needed","ask_index_status":"not_needed","reader_status":"not_needed","recognition_status":"not_needed","updated_at":"2026-05-14T23:06:18.309765+00:00","tier_text":"hub"},"tier":"hub","role_counts":[{"context_role":"background","n":1},{"context_role":"method","n":1}],"polarity_counts":[{"context_polarity":"background","n":1},{"context_polarity":"use_method","n":1}],"runs":{"context_extract":{"job_type":"context_extract","status":"succeeded","result":{"enqueued_papers":25},"error":null,"updated_at":"2026-05-14T06:57:36.804764+00:00"},"graph_features":{"job_type":"graph_features","status":"succeeded","result":{"co_cited":[{"title":"ReAct: Synergizing Reasoning and Acting in Language Models","work_id":"407a2351-25f1-497d-b611-f77d0292a8e6","shared_citers":25},{"title":"Self-Refine: Iterative Refinement with Self-Feedback","work_id":"59181e7f-e58e-45d3-8146-4477a9f53d5a","shared_citers":25},{"title":"Chain-of-Thought Prompting Elicits Reasoning in Large Language Models","work_id":"d1cf6693-a082-403c-ada9-dac7b96341f9","shared_citers":18},{"title":"Voyager: An Open-Ended Embodied Agent with Large Language Models","work_id":"ffe0d207-86cf-4742-a100-e988ac8b9676","shared_citers":18},{"title":"Evaluating Large Language Models Trained on Code","work_id":"042493e9-b26f-4b4e-bbde-382072ca9b08","shared_citers":16},{"title":"Toolformer: Language Models Can Teach Themselves to Use Tools","work_id":"9bce40c8-cfd7-4983-80e0-c3bd4402322a","shared_citers":16},{"title":"AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation","work_id":"92b7eb9c-c3d8-4518-a376-06fa15dd895b","shared_citers":15},{"title":"MemGPT: Towards LLMs as Operating Systems","work_id":"2698f5ad-c84c-40ca-b839-0912dae10ba2","shared_citers":14},{"title":"Training Verifiers to Solve Math Word Problems","work_id":"acab1aa8-b4d6-40e0-a3ee-25341701dca2","shared_citers":14},{"title":"Tree of Thoughts: Deliberate Problem Solving with Large Language Models","work_id":"07adb06e-4ed5-4ec5-a7ae-ff288fd214fb","shared_citers":14},{"title":"Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks","work_id":"27eaec54-c105-4969-8188-da5f0fca3688","shared_citers":13},{"title":"Self-Consistency Improves Chain of Thought Reasoning in Language Models","work_id":"8c6d5a6b-b5cc-4105-9c84-9c34bb9375bb","shared_citers":12},{"title":"GPT-4 Technical Report","work_id":"b928e041-6991-4c08-8c81-0359e4097c7b","shared_citers":11},{"title":"Training language models to follow instructions with human feedback","work_id":"52aff42f-4fa9-4fcf-bdb3-1459b9bebf65","shared_citers":11},{"title":"Constitutional AI: Harmlessness from AI Feedback","work_id":"faaaa4e0-2676-4fac-a0b4-99aef10d2095","shared_citers":10},{"title":"LLaMA: Open and Efficient Foundation Language Models","work_id":"c018fc23-6f3f-4035-9d02-28a2173b2b9d","shared_citers":9},{"title":"MetaGPT: Meta Programming for A Multi-Agent Collaborative Framework","work_id":"891b9780-a800-4e3c-bba0-53597ab8dc98","shared_citers":9},{"title":"DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models","work_id":"c5006563-f3ec-438a-9e35-b7b484f34828","shared_citers":8},{"title":"Qwen3 Technical Report","work_id":"25a4e30c-1232-48e7-9925-02fa12ba7c9e","shared_citers":8},{"title":"AgentBench: Evaluating LLMs as Agents","work_id":"a37549b4-4c94-412d-acc4-4efeb08509be","shared_citers":7},{"title":"DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning","work_id":"e6b75ad5-2877-4168-97c8-710407094d20","shared_citers":7},{"title":"Generative Agents: Interactive Simulacra of Human Behavior","work_id":"01f7ddaa-284a-441a-be87-921aad4dc54b","shared_citers":7},{"title":"Measuring Mathematical Problem Solving With the MATH Dataset","work_id":"50652ac6-fb7c-4675-a2c2-159c241feb17","shared_citers":7},{"title":"WebGPT: Browser-assisted question-answering with human feedback","work_id":"e25ef3e1-4848-4cb9-bf28-67a420591165","shared_citers":7}],"time_series":[{"n":9,"year":2023},{"n":2,"year":2024},{"n":69,"year":2026}],"dependency_candidates":[]},"error":null,"updated_at":"2026-05-14T07:07:25.905281+00:00"},"identity_refresh":{"job_type":"identity_refresh","status":"succeeded","result":{"items":[{"title":"Qwen3 Technical Report","outcome":"unchanged","work_id":"25a4e30c-1232-48e7-9925-02fa12ba7c9e","resolver":"local_arxiv","confidence":0.98,"old_work_id":"25a4e30c-1232-48e7-9925-02fa12ba7c9e"}],"counts":{"fixed":0,"merged":0,"unchanged":1,"quarantined":0,"needs_external_resolution":0},"errors":[],"attempted":1},"error":null,"updated_at":"2026-05-14T06:57:28.630506+00:00"},"summary_claims":{"job_type":"summary_claims","status":"succeeded","result":{"title":"Reflexion: Language Agents with Verbal Reinforcement Learning","claims":[{"claim_text":"Large language models (LLMs) have been increasingly used to interact with external environments (e.g., games, compilers, APIs) as goal-driven agents. However, it remains challenging for these language agents to quickly and efficiently learn from trial-and-error as traditional reinforcement learning methods require extensive training samples and expensive model fine-tuning. We propose Reflexion, a novel framework to reinforce language agents not by updating weights, but instead through linguistic feedback. Concretely, Reflexion agents verbally reflect on task feedback signals, then maintain the","claim_type":"abstract","evidence_strength":"source_metadata"}],"why_cited":"Pith tracks Reflexion: Language Agents with Verbal Reinforcement Learning because it crossed a citation-hub threshold.","role_counts":[]},"error":null,"updated_at":"2026-05-14T07:07:27.965753+00:00"}},"summary":{"title":"Reflexion: Language Agents with Verbal Reinforcement Learning","claims":[{"claim_text":"Large language models (LLMs) have been increasingly used to interact with external environments (e.g., games, compilers, APIs) as goal-driven agents. However, it remains challenging for these language agents to quickly and efficiently learn from trial-and-error as traditional reinforcement learning methods require extensive training samples and expensive model fine-tuning. We propose Reflexion, a novel framework to reinforce language agents not by updating weights, but instead through linguistic feedback. Concretely, Reflexion agents verbally reflect on task feedback signals, then maintain the","claim_type":"abstract","evidence_strength":"source_metadata"}],"why_cited":"Pith tracks Reflexion: Language Agents with Verbal Reinforcement Learning because it crossed a citation-hub threshold.","role_counts":[]},"graph":{"co_cited":[{"title":"ReAct: Synergizing Reasoning and Acting in Language Models","work_id":"407a2351-25f1-497d-b611-f77d0292a8e6","shared_citers":25},{"title":"Self-Refine: Iterative Refinement with Self-Feedback","work_id":"59181e7f-e58e-45d3-8146-4477a9f53d5a","shared_citers":25},{"title":"Chain-of-Thought Prompting Elicits Reasoning in Large Language Models","work_id":"d1cf6693-a082-403c-ada9-dac7b96341f9","shared_citers":18},{"title":"Voyager: An Open-Ended Embodied Agent with Large Language Models","work_id":"ffe0d207-86cf-4742-a100-e988ac8b9676","shared_citers":18},{"title":"Evaluating Large Language Models Trained on Code","work_id":"042493e9-b26f-4b4e-bbde-382072ca9b08","shared_citers":16},{"title":"Toolformer: Language Models Can Teach Themselves to Use Tools","work_id":"9bce40c8-cfd7-4983-80e0-c3bd4402322a","shared_citers":16},{"title":"AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation","work_id":"92b7eb9c-c3d8-4518-a376-06fa15dd895b","shared_citers":15},{"title":"MemGPT: Towards LLMs as Operating Systems","work_id":"2698f5ad-c84c-40ca-b839-0912dae10ba2","shared_citers":14},{"title":"Training Verifiers to Solve Math Word Problems","work_id":"acab1aa8-b4d6-40e0-a3ee-25341701dca2","shared_citers":14},{"title":"Tree of Thoughts: Deliberate Problem Solving with Large Language Models","work_id":"07adb06e-4ed5-4ec5-a7ae-ff288fd214fb","shared_citers":14},{"title":"Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks","work_id":"27eaec54-c105-4969-8188-da5f0fca3688","shared_citers":13},{"title":"Self-Consistency Improves Chain of Thought Reasoning in Language Models","work_id":"8c6d5a6b-b5cc-4105-9c84-9c34bb9375bb","shared_citers":12},{"title":"GPT-4 Technical Report","work_id":"b928e041-6991-4c08-8c81-0359e4097c7b","shared_citers":11},{"title":"Training language models to follow instructions with human feedback","work_id":"52aff42f-4fa9-4fcf-bdb3-1459b9bebf65","shared_citers":11},{"title":"Constitutional AI: Harmlessness from AI Feedback","work_id":"faaaa4e0-2676-4fac-a0b4-99aef10d2095","shared_citers":10},{"title":"LLaMA: Open and Efficient Foundation Language Models","work_id":"c018fc23-6f3f-4035-9d02-28a2173b2b9d","shared_citers":9},{"title":"MetaGPT: Meta Programming for A Multi-Agent Collaborative Framework","work_id":"891b9780-a800-4e3c-bba0-53597ab8dc98","shared_citers":9},{"title":"DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models","work_id":"c5006563-f3ec-438a-9e35-b7b484f34828","shared_citers":8},{"title":"Qwen3 Technical Report","work_id":"25a4e30c-1232-48e7-9925-02fa12ba7c9e","shared_citers":8},{"title":"AgentBench: Evaluating LLMs as Agents","work_id":"a37549b4-4c94-412d-acc4-4efeb08509be","shared_citers":7},{"title":"DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning","work_id":"e6b75ad5-2877-4168-97c8-710407094d20","shared_citers":7},{"title":"Generative Agents: Interactive Simulacra of Human Behavior","work_id":"01f7ddaa-284a-441a-be87-921aad4dc54b","shared_citers":7},{"title":"Measuring Mathematical Problem Solving With the MATH Dataset","work_id":"50652ac6-fb7c-4675-a2c2-159c241feb17","shared_citers":7},{"title":"WebGPT: Browser-assisted question-answering with human feedback","work_id":"e25ef3e1-4848-4cb9-bf28-67a420591165","shared_citers":7}],"time_series":[{"n":9,"year":2023},{"n":2,"year":2024},{"n":69,"year":2026}],"dependency_candidates":[]},"authors":[]}}