CoolFace
Apppublic

ai-sherpa/bayesian-stopping-consistent-llm-answers-repro

sourceHugging Faceupdated 2mo agoView on Hugging Face
0likes
3 commits on main
4462dff2mo ago

Claims 4-5: reproduce the paper's Section 5.2 query-count study on the official FEval-TTC cached LLM answers (36 cells, 17,055 replay trajectories, 0 live LLM calls)

ai-sherpa
8ac63462mo ago

Update logbook: Repro - Optimal Bayesian Stopping for Efficient Inference of Consistent LLM Answers

ai-sherpa
d8d94852mo ago

initial commit

ai-sherpa