Reproduce “Efficiently Deploying LLMs with Controlled Risk” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Michael J. Zellinger, Matt Thomson
HCMA uses calibrated correctness estimates from model token probabilities to let a query be answered by a small model, passed to a larger one, or rejected at any stage. On MMLU, it matches Llama3 405B selective-prediction performance at modeled average cost below $3 per million tokens, versus $5; this is not a live deployment comparison and requires labeled calibration data.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.