They're using the truncated SVD, not the full variant, that's computationally cheaper.
replyFair points, especially on GSM8K saturation and Qwen possibly already sitting close to the solution. That said, even if this is mostly "last-mile alignment", the fact that it can be done with such a tiny signal is still interesting, it suggests the gap between capability and behavior might be much smaller (and cheaper to bridge) than we assume.
replyYeah, my big problem with the paper is it just might be an artifact of qwen's training process.
replyIn all fairness most of the unique stuff I can do is probably an artifact of my training process, so it seems unfair to deny an LLM the same accomodation.
reply