corpus.blog
54,302 blogs · 4,481,250 posts Claim your blog
Record · corpus.blog/posts/01a08c59-68e3-7240-8151-0438b75bca4e

Recontextualization Mitigates Specification Gaming Without Modifying the Specification

turntrout.com · published 23 December 2025

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 turntrout.com Wayback capture
0blogs citing
First cited
—
Most recent
—
Rank this month
not ranked
Links from this post
35
Words captured
4,122

Cited by

No blog we hold has cited this post yet.

Links from this post

Link
Host
https://www.matsprogram.org/
matsprogram.org
https://arxiv.org/abs/2512.19027
arxiv.org
https://turntrout.com//team-shard
turntrout.com
https://www.lesswrong.com/posts/AXRHzCPMv6ywCxCFp/inoculation-prompting-instructing-models-to-misbehave-at
lesswrong.com
https://arxiv.org/abs/2510.05024
arxiv.org
https://arxiv.org/abs/2510.04340
arxiv.org
https://arxiv.org/abs/2209.15189
arxiv.org
https://arxiv.org/abs/2112.00861
arxiv.org
https://arxiv.org/pdf/2412.16339
arxiv.org
https://deepmind.google/discover/blog/specification-gaming-the-flip-side-of-ai-ingenuity/
deepmind.google
https://arxiv.org/abs/2310.13548
arxiv.org
https://arxiv.org/abs/2409.12822
arxiv.org
https://arxiv.org/abs/2503.11926
arxiv.org
https://metr.org/blog/2025-08-12-research-update-towards-reconciling-slowdown-with-time-horizons/
metr.org
https://www.lesswrong.com/posts/dbYEoG7jNZbeWX39o/training-a-reward-hacker-despite-perfect-labels
lesswrong.com
https://arxiv.org/abs/2508.17511
arxiv.org
https://www.lesswrong.com/editPost?postId=whkMnqFWKsBm7Gyd7&key=d961497becdfd870c17e9d586d0aa2
lesswrong.com
https://arxiv.org/abs/2108.07732
arxiv.org
https://www.lesswrong.com/posts/Ge55vxEmKXunFFwoe/reward-hacking-behavior-can-generalize-across-tasks
lesswrong.com
https://arxiv.org/abs/2505.13787
arxiv.org
https://huggingface.co/datasets/arianaazarbal/sycophancy_dataset
huggingface.co
https://huggingface.co/datasets/Anthropic/hh-rlhf/tree/main/helpful-base
huggingface.co
Deliberative alignment: reasoning enables safer language models Read at the source ↗
openai.com
https://www.apolloresearch.ai/research/stress-testing-anti-scheming-training
apolloresearch.ai
https://arxiv.org/abs/1707.01495
arxiv.org
https://forms.matsprogram.org/apply-s26
forms.matsprogram.org
https://github.com/arianaazarbal/recontextualization
github.com
https://arxiv.org/abs/2009.03300
arxiv.org
https://arxiv.org/abs/2311.07911
arxiv.org
https://arxiv.org/abs/2507.14805
arxiv.org
https://www.lesswrong.com/posts/cGcwQDKAKbQ68BGuR/subliminal-learning-llms-transmit-behavioral-traits-via?commentId=RifvqPmyiCKvKuXjc
lesswrong.com
https://arxiv.org/abs/2302.02676
arxiv.org
https://arxiv.org/abs/2302.05206
arxiv.org
https://arxiv.org/abs/2506.20061
arxiv.org
https://arxiv.org/abs/2407.16970
arxiv.org