Verifier strength does not buy discovery quality: a pre-registered two-domain test of LLM-based automated research, with post-hoc evidence pointing to proposer-pool composition
A post-hoc union-pool analysis shows the fastest path to better discovery is widening the pool, not tightening the gate: mixing the two pools adds +1.52 holdout-value units in domain A -- evidence that pool composition, not verification strength, is the lever.
Peter Chen, Ni Zhang, Si-Yuan Zhang
· 0 citations