conftrace_

reward modelling

2 papers

Explore in graph

Also known as

RM

Co-occurring keywords · papers mentioning both

trajectory label (1) robust generalisation (1) non-markovian reward (1) large language model (1) preference data (1) temporal dependency (1) interpretable model (1) human value (1) trajectory analysis (1) multiple instance learning (1)

Papers

Preference Learning for AI Alignment: a Causal Perspective ICML 2025 Non-Markovian Reward Modelling from Trajectory Labels via Interpretable Multiple Instance Learning NeurIPS 2022