The authors study direct in-context reinforcement learning, a controlled setting in which a model receives raw records of previous trajectories and rewards without parameter updates, memory retrieval or summarization.
Rewards Barely Drive Learning in Direct In-Context Reinforcement Learning
Across multiple tasks and models, trajectories improved performance largely regardless of whether their associated rewards were correct, randomized or absent.
Top University
Minchan Kwon · Seunghee Koh · Sunghyun Baek · Minsung Bae · Junmo Kim
Korea Advanced Institute of Science and Technology (KAIST)
Research Digest··2 min read
Kwon et al.
Why this paper
From Korea Advanced Institute of Science and Technology (KAIST)
In one line
Direct in-context reinforcement learning is driven by trajectory imitation and surface form, not by reward semantics, and should be reframed as in-context learning.
What we could check
- ·No code link found
- ·No weights link found
- ·No dataset link found
- ·No compute details found
- ✓Limitations stated by the authors (2 noted)
- ·No benchmark numbers found
Observed from the paper text and links we have. Absence here means we did not find it, not that it does not exist.
§