This paper proposes goal-conditioned supervised learning (GCSL) as an offline fine-tuning framework for LLMs and proposes natural-language goal representations to further connect these patterns to the LLM's pretrained knowledge and generalization capabilities.
RRCM is proposed, a ranking-driven retrieval-and-reasoning framework over collaborative and metadata memories for LLM-based agentic recommendation that significantly outperforms traditional baselines and diverse LLM-based recommendation approaches.
A Multi-Objective Goal-Conditioned Supervised Learning (MOGCSL) framework for automatically learning to achieve multiple objectives from offline sequential data and a novel goal-selection algorithm for MOGCSL to model and identify desired and achievable goals for inference.
Shi-Jun Li, Hilaf Hasson, Jing Hu et al.· arXiv.org· 1 citation
Self-play proposer--solver methods improve reasoning by generating tasks and learning from verified solutions. However, for evidence-identifiable tasks, where case-specific evidence and domain knowledge determine a checkable answer, self-play requires generating plausible cases whose answers can be independently verifi...