Record · corpus.blog/posts/01a0dca0-1d8d-710b-a9d6-82a5e43b72c0

Investigating the consequences of accidentally grading CoT during RL

alignment.openai.com · published 6 May 2026

A record is what was published and who pointed at it. The text of the post is not held here: read it at the source, or at its Wayback capture.

Read the post at alignment.openai.com Wayback capture

2blogs citing

First cited: 7 May 2026
Most recent: 20 May 2026
Rank this month: not ranked
Links from this post: 16
Words captured: 4,182

Cited by

Blog

In the post

Date

Last Week in AI

LWiAI Podcast #245 - TML-Interaction, Claude For Legal, Sam Altman on Stand Read at the source ↗

20 May 2026

Redwood Research

A review of “Investigating the consequences of accidentally grading CoT during RL” Read at the source ↗

7 May 2026

Links from this post

Link

Host

http://blog.redwoodresearch.org/openai-cot

blog.redwoodresearch.org

Detecting misbehavior in frontier reasoning models Read at the source ↗

openai.com

How we monitor internal coding agents for misalignment Read at the source ↗

openai.com

https://arxiv.org/abs/2507.11473

arxiv.org

Evaluating chain-of-thought monitorability Read at the source ↗

openai.com

https://www-cdn.anthropic.com/14e4fb01875d2a69f646fa5e574dea2b1c0ff7b5.pdf

www-cdn.anthropic.com

https://www-cdn.anthropic.com/3edfc1a7f947aa81841cf88305cb513f184c36ae.pdf

www-cdn.anthropic.com

https://deploymentsafety.openai.com/gpt-5-4-thinking/production-benchmarks-with-representative-prompts/fig-1

deploymentsafety.openai.com

https://arxiv.org/abs/2512.18311

arxiv.org

Where the goblins came from Read at the source ↗

openai.com

https://arxiv.org/abs/2603.30036

arxiv.org

https://alignment.openai.com/metagaming

alignment.openai.com

https://arxiv.org/abs/2503.11926

arxiv.org

https://arxiv.org/abs/2506.01926

arxiv.org

https://arxiv.org/abs/2511.11584

arxiv.org

https://alignment.anthropic.com/2025/subtle-reasoning/

alignment.anthropic.com