Thanh Q. Tran, John Russell Himawan, Arun Verma, Kiwan Wong, Wei Xiao, Daniela Rus, Bryan Kian Hsiang Low · 2026 · arXiv
DOI: 10.48550/arXiv.2610.31438 · open-access PDF
Quando um usuário diz "Mantenha-o curto", quão curto deve ser a resposta do LLM? Mais geralmente, usuários e aplicativos exigem diferentes níveis de adesão de instruções, mas o prompting não fornece confiável controle de grãos finos sobre eles. Portanto, estudamos "modelos de linguagem controláveis" que fornecem controle numérico de grãos finos sobre o grau em que os modelos seguem as instruções gerais especificadas no texto. No entanto, as abordagens existentes não oferecem conjuntamente controle de grãos finos, eficiência e escalabilidade em várias instruções. Para resolver esta lacuna, introduzimos o DIAL (Distilled Instruction-Adh
Machine translation (Meta M2M100); the English original follows.
When a user says "Keep it short", how short should the LLM's response be? More generally, users and applications require different levels of instruction adherence, but prompting does not reliably provide fine-grained control over them. We therefore study "steerable language models" that provide fine-grained numerical control over the degree to which models follow general instructions specified in text. However, existing approaches do not jointly offer fine-grained control, efficiency, and scalability across diverse instructions. To address this gap, we introduce DIAL (Distilled Instruction-Adh
Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese
GET https://store.agentexchange.work/gw/papers/verify?title=...&year=... (0.003 USD), fetch this record as JSON with /gw/papers/paper?id=2610.31438, or list what it cites with /gw/papers/references?id=.... Pay per call in USDC (x402) or with a 19 USD card key for 1,000 calls. Catalogue · llms.txtSources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange