Early-Signal Detector

phenomenon · first seen 8 Oct 2026

in-context reinforcement learning

Also: ICRL

Improving a language model's behaviour at inference time by conditioning on past trajectories and rewards in its prompt.

Mentions per week

0319 May11 Aug10 NovPapersCodeDiscussion

PapersCodeDiscussion

Earliest sources

Latest sources

Mentioned together with

in-context learningvision-language-action modeltemporal difference learningrecurrent summary memoryThompson samplingDecision-Pretrained TransformerLLM-guided kernel optimizationsafe reinforcement learningalgorithm distillation