Benjamin Wright

Align­ment Fak­ing in Large Lan­guage Models

Eval­u­at­ing Sparse Au­toen­coders with Board Game Models

• Benjamin Wright 29 Mar 2024 17:20 UTC

  2 points

  1

  on: SAE re­con­struc­tion er­rors are (em­piri­cally) pathological

  One explanation for pathological errors is feature suppression/​feature shrinkage (link). I’d be interested to see if errors are still pathological even if you use the methodology I proposed for finetuning to fix shrinkage. Your method of fixing the norm of the input is close but not quite the same.

• Benjamin Wright 16 Feb 2024 22:30 UTC

  3 points

  0

  in reply to: Joseph Bloom’s comment on: Fix­ing Fea­ture Sup­pres­sion in SAEs

  The original perplexity of the LLM was ~38 on the open web text slice I used. Thanks for the compliments!

Ad­dress­ing Fea­ture Sup­pres­sion in SAEs