It is suggested that multi-agent debate is most valuable when it preserves independent informational signals rather than simply improving measured reasoning quality, and that interventions enforcing stronger causal reasoning do not improve financial performance.
Ju-Li Huang, A. Alrassan, D. Harischandra et al.· 0 citations
Attention-head ablation, zeroing a head and measuring the resulting change in task performance, is a common method for inferring which components of a language model are causally responsible for a behavior. We show using GPT-2 small that this inference can be fragile unless the intervention semantics, evaluation metric...
A persistent agent must decide both what to retain as information arrives and what to surface once a query appears, yet memory evaluations can confound these decisions by comparing methods that differ in both retention and selection. We build a streaming-recall benchmark crossing retention and selection rules and evalu...
In domains with expert-designed symbolic features, preserving structure beats learning abstractions in domains with expert-designed symbolic features, and preserving structure beats learning abstractions in domains with domain-expert features.
Ju-Li Huang, Jake Cheng, Rupert Lu· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.