sylvain-wei/24-Game-Reasoning
超简单复现Deepseek-R1-Zero和Deepseek-R1，以「24点游戏」为例。通过zero-RL、SFT以及SFT+RL，以激发LLM的自主验证反思能力。 About Clean, minimal, accessible reproduction of DeepSeek R1-Zero, DeepSeek R1
This project helps AI researchers and developers improve how large language models (LLMs) solve mathematical problems. It takes an existing LLM and trains it using different techniques with the classic 24 Game. The output is a more accurate and robust LLM capable of better mathematical reasoning and self-correction, which can then be applied to other complex logical tasks.
No commits in the last 6 months.
Use this if you are developing or fine-tuning large language models and need to enhance their mathematical reasoning and self-verification abilities.
Not ideal if you are a casual user looking for a ready-to-play 24 Game solver or a general-purpose AI, as this is a research-focused development tool.
Stars
34
Forks
2
Language
Python
License
—
Category
Last pushed
Apr 05, 2025
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/llm-tools/sylvain-wei/24-Game-Reasoning"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
open-thought/reasoning-gym
[NeurIPS 2025 Spotlight] Reasoning Environments for Reinforcement Learning with Verifiable Rewards
Hmbown/Hegelion
Dialectical reasoning architecture for LLMs (Thesis → Antithesis → Synthesis)
LLM360/Reasoning360
A repo for open research on building large reasoning models
bowang-lab/BioReason
BioReason: Incentivizing Multimodal Biological Reasoning within a DNA-LLM Model | NeurIPS '25
TsinghuaC3I/Awesome-RL-for-LRMs
A Survey of Reinforcement Learning for Large Reasoning Models