Fine-tuned “student” models can pick up unwanted traits from base “teacher” models that could evade data filtering, generating a need for more rigorous safety evaluations. Researchers have discovered ...
A new study by Anthropic shows that language models might learn hidden characteristics during distillation, a popular method for fine-tuning models for special tasks. While these hidden traits, which ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results