Vai al contenuto
onsuggest
Accedi
Articolo en

Fun toy result: a subliminal-learning effect via on-policy distillation

Fonte: LessWrong ·

In the standard subliminal learning setup, a teacher model with a preference generates data according to its policy, and then the student is trained on those data. I wondered if it was necessary for the teacher model to actually get to control the produced tokens, so I instead let a no-prompt "teacher" model produce the sequences to be trained on. Then I trained the student against a subliminal-learning-prompted teacher model's logits on those sequences. (Since an unmodified "teacher" is identical to the student model, this is equivalent to just doing on-policy distillation from the prompted t

Che cos'è Fun toy result: a subliminal-learning effect via on-policy distillation?

Fun toy result: a subliminal-learning effect via on-policy distillation is a articolo discovery curated from LessWrong. In the standard subliminal learning setup, a teacher model with a preference generates data according to its policy, and then the student is trained on those data. I wondered if it

Come posso accedervi o provarlo?

You can experience this content directly on its original host (LessWrong) by clicking the "Open Original" button.

Da dove proviene questo contenuto?

Curated from LessWrong by the OnSuggest team to surface the most interesting and serendipitous content on the web.

Scorciatoie da tastiera

Scoperta precedente
←P
Scoperta successiva
→N
Mi piace
L
Non fa per me
X
Salva
S
Mostra questo aiuto
?
Chiudi
Esc

Le scorciatoie funzionano nella pagina delle scoperte.