microsoft/interwhen
A framework for verifiable reasoning with language models.
This project helps ensure the reliability and accuracy of language model outputs, especially for critical tasks where errors are unacceptable. It takes a prompt and a language model, and by continuously checking the model's intermediate thinking steps against specific rules, it either corrects, revises, or stops the generation process. This tool is designed for AI developers and engineers building high-stakes applications like those in law, healthcare, or robotics, where verifiable and correct AI reasoning is paramount.
Use this if you are developing language model-powered applications where individual errors are costly or dangerous, and you need to guarantee that outputs adhere to explicit, verifiable constraints.
Not ideal if your application prioritizes speed over absolute correctness, or if the domain lacks clear, verifiable rules to check against.
Stars
13
Forks
4
Language
Python
License
MIT
Category
Last pushed
Mar 19, 2026
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/transformers/microsoft/interwhen"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
ExtensityAI/symbolicai
A neurosymbolic perspective on LLMs
TIGER-AI-Lab/MMLU-Pro
The code and data for "MMLU-Pro: A More Robust and Challenging Multi-Task Language Understanding...
deep-symbolic-mathematics/LLM-SR
[ICLR 2025 Oral] This is the official repo for the paper "LLM-SR" on Scientific Equation...
zhudotexe/fanoutqa
Companion code for FanOutQA: Multi-Hop, Multi-Document Question Answering for Large Language...
xlang-ai/Binder
[ICLR 2023] Code for the paper "Binding Language Models in Symbolic Languages"