# Unakar/Logic-RL

Reproduce R1 Zero on Logic Puzzle

Repository: https://github.com/Unakar/Logic-RL
Canonical: https://ross.abutalabs.com/products/logic-rl
Language: Python
License: Apache-2.0
License Family: permissive
Last push: 2025-03-20T06:24:26+00:00

## Health v2 (maintenance only)
Score: 26/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 12, release rhythm 35, longevity 41
- inputs: {"age_days": 577, "days_push": 531, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2448, forks 163 (observed 2026-08-28T04:06:52.767715+00:00)

## What it is
Logic-RL is a research framework that reproduces R1-Zero-style LLM reasoning via rule-based reinforcement learning (GRPO) on logic puzzles, built on verl and vLLM. It includes data preprocessing, reward modeling, and training scripts for 4×A100 setups.

## Use cases
- train an LLM to reason with rule-based reinforcement learning
- reproduce R1 Zero emergent reasoning behavior
- run GRPO training on logic puzzle datasets
- benchmark small models on multi-person logic puzzles
- generate and preprocess knight-and-knave puzzle data

## When to choose
- you want to replicate or extend rule-based RL for LLM reasoning research
- you have multi-GPU (A100) resources and want a working verl/vLLM training pipeline
- you need a reproducible baseline for reasoning benchmarks

## When to avoid
- you need a production LLM serving or inference system
- you lack GPU cluster resources for RL training
- you want a general-purpose RL library rather than a research reproduction

## Facets
- artifact type: library
- maturity: active
- function: llm-training, reinforcement-learning, machine-learning, benchmarking
- domain: large-language-models, reinforcement-learning, machine-learning, artificial-intelligence
- platform: python
- tags: grpo, verl, vllm, rule-based-reward, logic-puzzles, reasoning, r1-zero, research-code, gpu, linux, docker

## Member repositories
- Unakar/Logic-RL (main) score 26

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:52.767715+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T02:30:27.527660+00:00, confidence not recorded.
  - readme: https://github.com/Unakar/Logic-RL (fetched 2026-08-28T04:06:52.767715+00:00, sha ab777ba05d7b)
- Data as of 2026-08-30T08:39:29.467469+00:00.