home/categories/machine-learning/openclaw-skills-skills-aiwithabidi-llm-evaluator-skill-md
machine-learningdata-ai

llm-evaluator

LLM-as-a-Judge evaluation system using Langfuse. Score AI outputs on relevance, accuracy, hallucination, and helpfulness. Backfill scoring on historical traces. Uses GPT-5-nano for cost-efficient judging. Use when evaluating AI quality, building evals, or monitoring output accuracy.

openclaw
maintainer
openclaw
Обновлено 3/5/2026
Звёзды
4001
Форки
1095
quick start

Installation and usage

LLM-as-a-Judge evaluation system using Langfuse. Score AI outputs on relevance, accuracy, hallucination, and helpfulness. Backfill scoring on historical traces. Uses GPT-5-nano for cost-efficient judging. Use when evaluating AI quality, building evals, or monitoring output accuracy.

Установка
$ install --globalskills.sh
Использование

После установки вы можете использовать этот skill, выполнив следующую команду в терминале:

skills use llm-evaluator