Early-Signal Detector

method · first seen 5 Oct 2026

reinforcement learning from task feedback

Post-training a model with reinforcement learning using rewards from task success signals, such as passing tests.

Mentions per week

0119 May11 Aug10 NovPapersCodeDiscussion

PapersCodeDiscussion

Earliest sources

Latest sources

Mentioned together with

backdoor attacksupervised fine-tuning