Usage
from transformers import AutoModelForCausalLM, AutoTokenizer
tok = AutoTokenizer.from_pretrained("pb09204048/CRISP-Qwen3-14B-v1")
model = AutoModelForCausalLM.from_pretrained("pb09204048/CRISP-Qwen3-14B-v1", device_map="auto")
Benchmark results (Qwen3-14B)
Accuracy (mean@8, %) and token reduction (Red., % vs. base) at a 30K-token budget. Math is scored with
a dual-path grader (Answer: or \boxed{}); GPQA-Diamond and MMLU use exact letter-match. This
model is the CRISP (v1) row.
Table with columns: Setting, MATH-500, AIME 2024, AIME 2025, GPQA-D, MMLU| Setting | MATH-500 | AIME 2024 | AIME 2025 | GPQA-D | MMLU |
|---|
| Base | 93.0 / — | 75.0 / — | 69.2 / — | 62.2 / — | 85.1 / — |
| Concise prompt (v2) | 94.3 / 25.7% | 73.3 / 13.0% | 71.7 / 10.1% | 60.5 / 26.0% | 84.9 / 22.2% |
| Concise prompt (v1) | 95.9 / 43.1% | 76.7 / 23.5% | 66.2 / 20.1% | 60.6 / 26.1% | 84.9 / 21.8% |
| CRISP (v2) | 95.2 / 34.7% | 75.0 / 19.7% | 67.1 / 16.8% | 62.0 / 20.7% | 83.9 / 22.4% |
| CRISP (v1) | 96.3 / 56.3% | 73.8 / 37.5% | 62.9 / 32.1% | 61.9 / 39.7% | 84.2 / 43.1% |
Citation
@article{sang2026crisp,
title={Crisp: Compressed reasoning via iterative self-policy distillation},
author={Sang, Hejian and Xu, Yuanda and Zhou, Zhengze and He, Ran and Wang, Zhipeng and Sun, Jiachen},
journal={arXiv preprint arXiv:2603.05433},
year={2026}
}