Blog
LLMs & Texto
The Curse of Multiple Mediators: Hidden Interaction Effects in Activation Patching
arXiv:2606.27510v1 Announce Type: new Abstract: Activation patching is the primary tool in mechanistic interpretability. It attributes causal responsibility for a model behavior to each of its individual components by estimating its natural indirect effect (NIE). Re-deriving the activation patching estimand from causal mediation analysis, we find that the NIE does not solely capture the causal effect through the specific component. It also contains interaction effects (INT) that measure how much...
arXiv cs.LG
·Sankaran Vaidyanathan, David Arbour, Aaron Mueller, Scott Niekum, David Jensen
·
// relacionados
Leia também
Blog
Orquestração agêntica: as organizações de IA corporativa têm um problema de implantação, não um problema de plataforma — e a maioria está chamando chatbots de agentes
Blog
Soofi Consortium lança o Soofi S 30B-A3B: um modelo de fundação MoE híbrido Mamba-Transformer aberto para alemão e inglês
Modelo
thinkingmachines/Inkling
Blog