Global foundation-model progress briefing
English Edition中文
Enter keywords to search ingested stories.

Today / Friday, November 20, 2026

limbo logolimbo

Data updated

Nov 9, 06:52 PM

Live sources

17

Ingestion status

Live ingest

PolicyarXiv AI / CL

Shortcuts and Boundaries: Verifiable RL Misalignment and Contextual Safety Learning

Summary

Reinforcement learning (RL) can improve an agent's task capability while reinforcing behavior that violates contextual execution requirements. We study this reward-driven misalignment, including reward hacking, through verifiable reward--context conflicts, whe...

Original Article

Captured source content or English translation, normalized into this reading format.

Read Source
This story does not yet have captured source text. Open the source link to read it.

Region

Global

Heat Score

73

Category

Policy

Language

en