Jannik Brinkmann

Eval­u­at­ing Sparse Au­toen­coders with Board Game Models

In­ter­pret­ing Prefer­ence Models w/​ Sparse Autoencoders

Find­ing Back­ward Chain­ing Cir­cuits in Trans­form­ers Trained on Tree Search

Im­prov­ing SAE’s by Sqrt()-ing L1 & Re­mov­ing Low­est Ac­ti­vat­ing Fea­tures