Matches in SemOpenAlex for { <https://semopenalex.org/work/W4387560542> ?p ?o ?g. }
Showing items 1 to 71 of
71
with 100 items per page.
- W4387560542 abstract "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 backbone LMs. For example, fine-tuning Llama2-7B with 500 agent trajectories generated by GPT-4 leads to a 77% HotpotQA performance increase. Furthermore, we propose FireAct, a novel approach to fine-tuning LMs with trajectories from multiple tasks and prompting methods, and show having more diverse fine-tuning data can further improve agents. Along with other findings regarding scaling effects, robustness, generalization, efficiency and cost, our work establishes comprehensive benefits of fine-tuning LMs for agents, and provides an initial set of experimental designs, insights, as well as open questions toward language agent fine-tuning." @default.
- W4387560542 created "2023-10-12" @default.
- W4387560542 creator A5000577985 @default.
- W4387560542 creator A5014888975 @default.
- W4387560542 creator A5025205227 @default.
- W4387560542 creator A5053921484 @default.
- W4387560542 creator A5073413742 @default.
- W4387560542 creator A5086032589 @default.
- W4387560542 date "2023-10-09" @default.
- W4387560542 modified "2023-10-18" @default.
- W4387560542 title "FireAct: Toward Language Agent Fine-tuning" @default.
- W4387560542 doi "https://doi.org/10.48550/arxiv.2310.05915" @default.
- W4387560542 hasPublicationYear "2023" @default.
- W4387560542 type Work @default.
- W4387560542 citedByCount "0" @default.
- W4387560542 crossrefType "posted-content" @default.
- W4387560542 hasAuthorship W4387560542A5000577985 @default.
- W4387560542 hasAuthorship W4387560542A5014888975 @default.
- W4387560542 hasAuthorship W4387560542A5025205227 @default.
- W4387560542 hasAuthorship W4387560542A5053921484 @default.
- W4387560542 hasAuthorship W4387560542A5073413742 @default.
- W4387560542 hasAuthorship W4387560542A5086032589 @default.
- W4387560542 hasBestOaLocation W43875605421 @default.
- W4387560542 hasConcept C104317684 @default.
- W4387560542 hasConcept C121332964 @default.
- W4387560542 hasConcept C134306372 @default.
- W4387560542 hasConcept C136197465 @default.
- W4387560542 hasConcept C137293760 @default.
- W4387560542 hasConcept C154945302 @default.
- W4387560542 hasConcept C157524613 @default.
- W4387560542 hasConcept C177148314 @default.
- W4387560542 hasConcept C177264268 @default.
- W4387560542 hasConcept C185592680 @default.
- W4387560542 hasConcept C199360897 @default.
- W4387560542 hasConcept C33923547 @default.
- W4387560542 hasConcept C41008148 @default.
- W4387560542 hasConcept C55493867 @default.
- W4387560542 hasConcept C62520636 @default.
- W4387560542 hasConcept C63479239 @default.
- W4387560542 hasConceptScore W4387560542C104317684 @default.
- W4387560542 hasConceptScore W4387560542C121332964 @default.
- W4387560542 hasConceptScore W4387560542C134306372 @default.
- W4387560542 hasConceptScore W4387560542C136197465 @default.
- W4387560542 hasConceptScore W4387560542C137293760 @default.
- W4387560542 hasConceptScore W4387560542C154945302 @default.
- W4387560542 hasConceptScore W4387560542C157524613 @default.
- W4387560542 hasConceptScore W4387560542C177148314 @default.
- W4387560542 hasConceptScore W4387560542C177264268 @default.
- W4387560542 hasConceptScore W4387560542C185592680 @default.
- W4387560542 hasConceptScore W4387560542C199360897 @default.
- W4387560542 hasConceptScore W4387560542C33923547 @default.
- W4387560542 hasConceptScore W4387560542C41008148 @default.
- W4387560542 hasConceptScore W4387560542C55493867 @default.
- W4387560542 hasConceptScore W4387560542C62520636 @default.
- W4387560542 hasConceptScore W4387560542C63479239 @default.
- W4387560542 hasLocation W43875605421 @default.
- W4387560542 hasOpenAccess W4387560542 @default.
- W4387560542 hasPrimaryLocation W43875605421 @default.
- W4387560542 hasRelatedWork W2032233321 @default.
- W4387560542 hasRelatedWork W2899852118 @default.
- W4387560542 hasRelatedWork W3023285645 @default.
- W4387560542 hasRelatedWork W3023594376 @default.
- W4387560542 hasRelatedWork W3154646238 @default.
- W4387560542 hasRelatedWork W4287802662 @default.
- W4387560542 hasRelatedWork W4309877123 @default.
- W4387560542 hasRelatedWork W4362554255 @default.
- W4387560542 hasRelatedWork W4378468482 @default.
- W4387560542 hasRelatedWork W3037551068 @default.
- W4387560542 isParatext "false" @default.
- W4387560542 isRetracted "false" @default.
- W4387560542 workType "article" @default.