Read the piece on lilianweng.github.io ↗
Every claim below was made in this piece, quoted word for word and numbered in the order the piece makes them, so you can read it there rather than take our word for it. The sentence above each quote is our reading of the claim, not their wording. Each quote was checked against a stored copy of the page at build time; where the two differ, the quote is the fact.
Our reading
Large language models cannot reliably self-correct their own mistakes without external feedback.
Mental modelError correction
Their wordsthis self-correction capability turns out to not exist intrinsically among LLMs and does not easily work out of the box, due to various failure modes
Our reading
Longer test-time thinking improves an AI model's robustness to adversarial or unusual inputs.
Their wordsthinking for longer should be especially useful when the model is presented with an unusual input, such as an adversarial example or jailbreak attempt
Our reading
Mental modelIncentives
Their wordsmodel CoTs could be biased due to lack of explicit training objectives aimed at encouraging faithful reasoning
Keep reading
The other pieces are the ones whose claims are worded most like this one's. That measures language, not agreement: they may argue the opposite.