Decoding Adjustment
A knowledge extraction enhancement technique that modifies the token selection strategy during generation without touching model weights or internal representations. Representative methods include contrastive decoding and DoLa.
Timeline (last 90 days)
Knowledge extraction enhancement (KEE) techniques include three categories—activation steering, decoding adjustments, and prompt engineering—which modify the inference process and may reinforce existing predictions or surface answers that are otherwise hard to reach.
Decoding adjustments do not touch model weights or internal representations, but instead modify the token selection strategy during the generation stage; contrastive decoding runs a strong model and a weak model simultaneously and takes the difference in their log probabilities as the score for the next token; DoLa compares the output distributions of different layers within a single model.
All Facts (2)
Knowledge extraction enhancement (KEE) techniques include three categories—activation steering, decoding adjustments, and prompt engineering—which modify the inference process and may reinforce existing predictions or surface answers that are otherwise hard to reach.
50%UnverifiedDecoding adjustments do not touch model weights or internal representations, but instead modify the token selection strategy during the generation stage; contrastive decoding runs a strong model and a weak model simultaneously and takes the difference in their log probabilities as the score for the next token; DoLa compares the output distributions of different layers within a single model.
50%