AIToday
AI Safety & AlignmentLessWrong AIPublished: Nov 16, 2026, 06:00 JST

Alignment 'train/test splits' should be sacred, essay argues

Send on LINE
Alignment 'train/test splits' should be sacred, essay argues

A LessWrong essay makes the case that alignment techniques can be used for training alignment or for testing it, but not both, and wants some held out only for evaluation.

Not sure about something? Ask the AI

Questions and answers are published on this page.

Summaries like this, every morning.

Send on LINE

AI news that matters for your work, delivered every morning.

Pick your industry and the AI tools you use, and get news related to your work every day.

Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.

Questions and answers are published on this page.

Related Articles

Next articleSingle </think> weight row makes reasoning models stop early, no training