We trained 13,440 large language models and found that entropy minimization requires only a single unlabeled data and 10 steps optimization to achieve performance improvements comparable to or even greater than those obtained using thousands of data and carefully designed rewards in rule-based reinforcement learning. This striking result may prompt a rethinking of post-training paradigms for large language models.

Code: https://github.com/zitian-gao/one-shot-em Project Page: https://www.notion.so/One-shot-Entropy-Minimization-202606db813b80639773f850f39246a5

Installation

pip install torch transformers==4.47.1 accelerate deepspeed psutil pandas numpy wandb

Reproducing One-shot EM Training (SOTA)

bash

accelerate launch train.py \
  --model_name Qwen2.5-Math-7B \
  --model_path /path/to/Qwen2.5-Math-7B \
  --train_data dataset/1shot_rlvr/pi1_r1280.parquet \
  --effective_batch 64 \
  --micro_batch_size 2 \
  --temperature 0.5 \
  --learning_rate 2e-5 \
  --max_steps 50 \
  --log_steps 1 \
  --save_steps 1 \
  --run_name one_shot \
  --wandb_project one-shot-em

Reproducing Multi-shot EM Training

bash

accelerate launch train.py \
  --model_name Qwen2.5-Math-7B \
  --model_path /path/to/Qwen2.5-Math-7B \
  --train_data dataset/numina/numina_00.parquet \
  --effective_batch 64 \
  --micro_batch_size 2 \
  --temperature 0.5 \
  --learning_rate 2e-5 \
  --max_steps 50 \
  --log_steps 1 \
  --save_steps 1 \
  --run_name multi_shot \
  --wandb_project one-shot-em

Evaluation

bash

cd Qwen2.5-Eval/evaluation
bash sh/eval_all_math.sh

Acknowledgements

Our dataset references and builds upon the following open-source contributions:

NuminaMath-CoT
DeepScaler
One-shot RLVR – for data selection strategies
Qwen2.5-Eval – for evaluation benchmarks

We sincerely thank the authors and maintainers of these projects for their excellent contributions to the research community!

Citation

@misc{gao2025oneshotentropyminimization,
      title={One-shot Entropy Minimization}, 
      author={Zitian Gao and Lynx Chen and Joey Zhou and Bryan Dai},
      year={2025},
      eprint={2505.20282},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2505.20282}, 
}

zgao3186/qwen25math7b-one-shot-em

One-shot Entropy Minimization

Installation

Reproducing One-shot EM Training (SOTA)

Reproducing Multi-shot EM Training

Evaluation

Acknowledgements

Citation

zgao3186/
qwen25math7b-one-shot-em