Tag: HuggingFace

Showing 1-2 of 2

Jun 25, 20267 min readTooling Deep DiveIntermediate180 min build

Olmo Hybrid vs Olmo 3 — which token types each model predicts better

Reproducible token-level tests comparing Olmo Hybrid and Olmo 3 show hybrids better on meaning-bearing tokens (nouns, verbs, adjectives, coref), transformers win on verbatim copy.

hybrid-models transformers token-evaluation Olmo-Hybrid Olmo-3

+3 more

TutorialsFrance

Open

Jun 21, 20267 min readAgent PlaybookIntermediate240 min build

Measuring how open models use your libraries: a reproducible agent benchmark

Build a repeatable harness that records agents' plan steps, API calls, retries, tokens, wall time and cost to reveal friction points in your library and guide rollout decisions.

agents benchmarking open-models tooling huggingface

+3 more

evaluation pi-agent observability