DolbyUUU/DeepEnlighten
Pure RL to post-train base models for social reasoning capabilities. Lightweight replication of DeepSeek-R1-Zero with Social IQa dataset.
This project helps AI researchers and developers working on large language models (LLMs) improve their models' social reasoning abilities. By applying pure reinforcement learning to existing base models like Llama or Qwen, it takes a base model and the Social IQa dataset as input and outputs a post-trained model with enhanced social intelligence. This is ideal for those looking to explore how social reasoning generalizes to other cognitive tasks.
No commits in the last 6 months.
Use this if you are an AI researcher or developer experimenting with reinforcement learning to enhance the emotional and social intelligence of base language models.
Not ideal if you are looking for a plug-and-play solution for end-user applications or if your primary goal is to train models for tasks unrelated to social or mathematical reasoning.
Stars
39
Forks
—
Language
Python
License
—
Category
Last pushed
Mar 16, 2025
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/transformers/DolbyUUU/DeepEnlighten"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
agentscope-ai/Trinity-RFT
Trinity-RFT is a general-purpose, flexible and scalable framework designed for reinforcement...
OpenRLHF/OpenRLHF
An Easy-to-use, Scalable and High-performance Agentic RL Framework based on Ray (PPO & DAPO &...
zjunlp/EasyEdit
[ACL 2024] An Easy-to-use Knowledge Editing Framework for LLMs.
huggingface/alignment-handbook
Robust recipes to align language models with human and AI preferences
hyunwoongko/nanoRLHF
nanoRLHF: from-scratch journey into how LLMs and RLHF really work.