3 ms·
Why they use different models to decode the reasoning content? Can the the model decode it?
by C0ldSmi1e 2mo ago
Why they use different models to decode the reasoning content? Can the the model decode it?
- aszen 2mo agoBecause stronger models are harder to jailbreak from the paper it says haiku was easily fooled into giving us thinking contents