giang_at_glai

4 ms·

1. 1.▲
   by giang_at_glai 7mo ago

   Actually, the model is forcing the response to be generated inside the attribution modules.
2. 2.▲
   by giang_at_glai 7mo ago

   We will share a technical write-up soon that addresses both of your questions: (1) steering vs. prompt engineering, and (2) how effectively our steering suppresses undesired generations. If you have joined our waitlist, we will notify you a
3. 3.▲
   by giang_at_glai 7mo ago

   Author here. This post shows “concept algebra” on language model: inject, suppress, and compose human-understandable concepts at inference time (no retraining, no prompt engineering). There’s an interactive demo on the post. Would love feed