all AI news
Dialogue Action Tokens: Steering Language Models in Goal-Directed Dialogue with a Multi-Turn Planner
June 19, 2024, 4:40 a.m. | Kenneth Li, Yiming Wang, Fernanda Vi\'egas, Martin Wattenberg
cs.CL updates on arXiv.org arxiv.org
Abstract: We present an approach called Dialogue Action Tokens (DAT) that adapts language model agents to plan goal-directed dialogues. The core idea is to treat each utterance as an action, thereby converting dialogues into games where existing approaches such as reinforcement learning can be applied. Specifically, we freeze a pretrained language model and train a small planner model that predicts a continuous action vector, used for controlled generation in each round. This design avoids the problem …
abstract action agents arxiv core cs.ai cs.cl cs.lg dat dialogue games language language model language models reinforcement reinforcement learning steering tokens type
More from arxiv.org / cs.CL updates on arXiv.org
Jobs in AI, ML, Big Data
AI Focused Biochemistry Postdoctoral Fellow
@ Lawrence Berkeley National Lab | Berkeley, CA
Senior Data Engineer
@ Displate | Warsaw
PhD Student AI simulation electric drive (f/m/d)
@ Volkswagen Group | Kassel, DE, 34123
AI Privacy Research Lead
@ Leidos | 6314 Remote/Teleworker US
Senior Platform System Architect, Silicon
@ Google | New Taipei, Banqiao District, New Taipei City, Taiwan
Fabrication Hardware Litho Engineer, Quantum AI
@ Google | Goleta, CA, USA