semantic-entropy-gate implements the semantic entropy method from Farquhar et al. (Nature 2024) to quantify uncertainty and detect confabulation in large language models. It can be used to gate actions in LLM-based agents, reducing the risk of acting on hallucinated information. The library is designed for integration into autonomous agent frameworks and is available as an open-source Python package.
In the LLM eval & observability space, semantic-entropy-gate takes a focused approach. It focuses on detecting and preventing LLM confabulation or hallucination in autonomous agents before they take actions. It is built as an open-source project for AI agent developers. semantic-entropy-gate is open source under the MIT license. semantic-entropy-gate is available on the command line and API.
krishddd builds and maintains semantic-entropy-gate, and the product first shipped in 2026. Development happens publicly on GitHub with 15 commits in the last 90 days. Key capabilities include Semantic Entropy, Hallucination Detection, and Agent Guardrails.
Latest indexed changes and source events
semantic-entropy-gate verified by the PulseGate indexer
Other apps tracked under the same category.
