@article{zhang2025dualspace, title = {A Dual-Space Framework for General Knowledge Distillation of Large Language Models}, author = {Zhang, Xue and Zhang, Songming and Liang, Yunlong and Meng, Fandong and Chen, Yufeng and Xu, Jinan and Zhou, Jie}, journal = {arXiv preprint arXiv:2504.11426}, year = {2025}, eprint = {2504.11426}, archiveprefix = {arXiv}, primaryclass = {cs.CL}, url = {https://arxiv.org/abs/2504.11426} } @article{ye2025gad, title = {Black-Box On-Policy Distillation of Large Language Models}, author = {Ye, Tianzhu and Dong, Li and Chi, Zewen and Wu, Xun and Huang, Shaohan and Wei, Furu}, journal = {arXiv preprint arXiv:2511.10643}, year = {2025}, eprint = {2511.10643}, archiveprefix = {arXiv}, primaryclass = {cs.CL}, url = {https://arxiv.org/abs/2511.10643} } @inproceedings{ramakrishnan2025omnidraft, title = {{OmniDraft}: A Cross-vocabulary, Online Adaptive Drafter for On-device Speculative Decoding}, author = {Ramakrishnan, Ramchalam Kinattinkara and Yuan, Zhaocong and Zhuo, Shaojie and Feng, Chen and Lin, Yicheng and Su, Chenzheng and Zhang, Xiaopeng}, booktitle = {Advances in Neural Information Processing Systems}, year = {2025}, url = {https://arxiv.org/abs/2507.02659} } @misc{ko2026reopold, title = {Relaxed On-Policy Distillation: Selective Credit Allocation for Scaling Reasoning Efficiently}, author = {Ko, Jongwoo and Abdali, Sara and Kim, Young Jin and Chen, Tianyi and Cameron, Pashmina}, howpublished = {Author publication page lists NeurIPS 2026}, year = {2026}, url = {https://www.microsoft.com/en-us/research/publication/relaxed-on-policy-distillation-selective-credit-allocation-for-scaling-reasoning-efficiently/}, note = {arXiv:2603.11137 v1 has an earlier title; the latest author-linked OpenReview full text was not retrieved at the 2026-10-14 refresh} } @article{xiong2026ovd, title = {{OVD}: On-policy Verbal Distillation}, author = {Xiong, Jing and Shen, Hui and Gong, Shansan and Cheng, Yuxin and Shen, Jianghan and Tao, Chaofan and Tan, Haochen and Bai, Haoli and Shang, Lifeng and Wong, Ngai}, journal = {arXiv preprint arXiv:2601.21968}, year = {2026}, eprint = {2601.21968}, archiveprefix = {arXiv}, primaryclass = {cs.CL}, url = {https://arxiv.org/abs/2601.21968}, note = {Technical report} } @article{zhang2026kdflow, title = {{KDFlow}: A User-Friendly and Efficient Knowledge Distillation Framework for Large Language Models}, author = {Zhang, Songming and Zhang, Xue and Zhang, Tong and Hu, Bojie and Chen, Yufeng and Xu, Jinan}, journal = {arXiv preprint arXiv:2603.01875}, year = {2026}, eprint = {2603.01875}, archiveprefix = {arXiv}, primaryclass = {cs.CL}, url = {https://arxiv.org/abs/2603.01875} } @article{sun2026simct, title = {{SimCT}: Recovering Lost Supervision for Cross-Tokenizer On-Policy Distillation}, author = {Sun, Jie and Zheng, Mao and Song, Mingyang and Zhong, Qiyong and Cheng, Yilin and Feng, Bichuan and Liu, Pengfei and Fang, Junfeng and Wang, Xiang}, journal = {arXiv preprint arXiv:2605.07711}, year = {2026}, eprint = {2605.07711}, archiveprefix = {arXiv}, primaryclass = {cs.CL}, url = {https://arxiv.org/abs/2605.07711} } @article{fang2026ropd, title = {Rubric-based On-policy Distillation}, author = {Fang, Junfeng and Hong, Zhepei and Zheng, Mao and Song, Mingyang and Li, Gengsheng and Jiang, Houcheng and Zhang, Dan and Guo, Haiyun and Wang, Xiang and Chua, Tat-Seng}, journal = {arXiv preprint arXiv:2605.07396}, year = {2026}, eprint = {2605.07396}, archiveprefix = {arXiv}, primaryclass = {cs.LG}, url = {https://arxiv.org/abs/2605.07396} } @article{zhou2026esr, title = {Less is More: Early Stopping Rollout for On-Policy Distillation}, author = {{Zhou Ziheng} and Li, Jiaqi and Tang, Huacong and Wu, Ying Nian and Terzopoulos, Demetri}, journal = {arXiv preprint arXiv:2605.27028}, year = {2026}, eprint = {2605.27028}, archiveprefix = {arXiv}, primaryclass = {cs.LG}, url = {https://arxiv.org/abs/2605.27028} } @article{zhou2026omniopd, title = {{OmniOPD}: Logit-Free On-Policy Distillation via Speculative Verification}, author = {Zhou, Yuhang and Zhang, Lizhu and Wu, Yifan and Wang, Mingyi and Peng, Bo and Liu, Jiayi and Fan, Xiangjun and Zhao, Zhuokai}, journal = {arXiv preprint arXiv:2606.01476}, year = {2026}, eprint = {2606.01476}, archiveprefix = {arXiv}, primaryclass = {cs.LG}, url = {https://arxiv.org/abs/2606.01476} } @article{yang2026oprd, title = {{OPRD}: On-Policy Representation Distillation}, author = {Yang, Shenzhi and Zhu, Guangcheng and Song, Bowen and Wang, Haobo and Xia, Mingxuan and Zheng, Xing and Ma, Yingfan and Chen, Zhongqi and Wang, Weiqiang and Zhao, Junbo and Chen, Gang}, journal = {arXiv preprint arXiv:2606.06021}, year = {2026}, eprint = {2606.06021}, archiveprefix = {arXiv}, primaryclass = {cs.LG}, url = {https://arxiv.org/abs/2606.06021} } @article{niu2026btb, title = {Breaking the Tokenizer Barrier: On-Policy Distillation across Model Families}, author = {Niu, Yifan and Xiao, Han and Liu, Dongyi and Wang, Zelong and Gong, Dihong and Wang, Yasheng and Li, Jia}, journal = {arXiv preprint arXiv:2606.09456}, year = {2026}, eprint = {2606.09456}, archiveprefix = {arXiv}, primaryclass = {cs.LG}, url = {https://arxiv.org/abs/2606.09456} } @article{sun2026easyopd, title = {{EasyOPD}: An Easy-to-use On-Policy Distillation Framework for Large Language Models}, author = {Sun, Jie and Zheng, Mao and Song, Mingyang and Zhong, Qiyong and Li, Gengsheng and Hong, Zhepei and Wu, Chang and Liu, Pengfei and Fang, Junfeng and Wang, Xiang}, journal = {arXiv preprint arXiv:2607.11012}, year = {2026}, eprint = {2607.11012}, archiveprefix = {arXiv}, primaryclass = {cs.CL}, url = {https://arxiv.org/abs/2607.11012} } @article{wang2026bpm, title = {Cross-Tokenizer On-Policy Distillation via Byte-Prefix Marginalization}, author = {Wang, Hao and Yuan, Kun and Zhong, Wenlin and Zhang, Minglei and Xiao, Han and Sun, Ming and Qi, Honggang}, journal = {arXiv preprint arXiv:2607.22334}, year = {2026}, eprint = {2607.22334}, archiveprefix = {arXiv}, primaryclass = {cs.LG}, url = {https://arxiv.org/abs/2607.22334} } @misc{patino2025gold, title = {Unlocking On-Policy Distillation for Any Model Family}, author = {Patiño, Carlos Miguel and Rasul, Kashif and Gallouédec, Quentin and Burtenshaw, Ben and Paniego, Sergio and Srivastav, Vaibhav and Frere, Thibaud and Beeching, Ed and Tunstall, Lewis and von Werra, Leandro and Wolf, Thomas}, year = {2025}, howpublished = {Hugging Face research article and TRL implementation}, url = {https://huggingface.co/spaces/HuggingFaceH4/on-policy-distillation}, note = {Non-paper resource} } @article{he2026simpleopd, title = {SimpleOPD: Simple Tokenizer-Agnostic On-Policy Distillation for Long-Context Reasoning}, author = {He, Haonan and Lei, Haodi and Luo, Yun and Zhang, Haoran and Zhang, Shunkai and Li, Yizhuo and Tang, Shengji and Wang, Zhilin and Zhan, Runzhe and Bai, Lei and Cui, Ganqu and Yu, Fangchen and Li, Yafu and Ye, Peng and Ding, Ning and Cheng, Yu}, journal = {arXiv preprint arXiv:2608.14277}, year = {2026}, eprint = {2608.14277}, archiveprefix = {arXiv}, primaryclass = {cs.CL}, url = {https://arxiv.org/abs/2608.14277} } @article{gu2026compassopd, title = {CompassOPD: Cross-Family On-Policy Distillation via Within-Family Likelihood Shifts}, author = {Gu, Naibin and Si, Qingyi and Yang, Chenxu and Qin, Chuanyu and Zhou, Junhao and Fu, Peng and Lin, Zheng and Wang, Weiping}, journal = {arXiv preprint arXiv:2609.10154}, year = {2026}, eprint = {2609.10154}, archiveprefix = {arXiv}, primaryclass = {cs.LG}, url = {https://arxiv.org/abs/2609.10154} } @article{chen2026dvlm, title = {Dual-Vocabulary Language Model for Cross-Tokenizer Distillation}, author = {Chen, Kedi and Lin, Chen and Sun, Yutao and Zhang, Wei}, journal = {arXiv preprint arXiv:2609.33816}, year = {2026}, eprint = {2609.33816}, archiveprefix = {arXiv}, primaryclass = {cs.AI}, url = {https://arxiv.org/abs/2609.33816} } @article{liu2026escd, title = {Beyond Token Alignment: Event Completion for Cross-Tokenizer On-Policy Distillation}, author = {Liu, Jiacheng and Song, Jingwei and Zhang, Qituan and Chen, Siheng and Zhang, Linfeng}, journal = {arXiv preprint arXiv:2609.34738}, year = {2026}, eprint = {2609.34738}, archiveprefix = {arXiv}, primaryclass = {cs.CL}, url = {https://arxiv.org/abs/2609.34738} } @article{hou2026rethinking, title = {Rethinking Cross-Tokenizer On-Policy Distillation: From Alignment Coverage to Supervision Reliability}, author = {Hou, Bingxi and Jiang, Guochao and Quan, Guofeng and Li, Weiqing and Feng, Wenfeng and Liu, Guohua and Zhang, Yuewei}, journal = {arXiv preprint arXiv:2610.08448}, year = {2026}, eprint = {2610.08448}, archiveprefix = {arXiv}, primaryclass = {cs.CL}, url = {https://arxiv.org/abs/2610.08448} } @article{yang2026cast, title = {CAST: Credit Alignment for Cross-Tokenizer On-Policy Distillation}, author = {Yang, Jingru and Wu, Jinhang and Shi, Minjing and He, Shengfeng and Li, Xuyan and Tan, Man and Li, Biye}, journal = {arXiv preprint arXiv:2610.13257}, year = {2026}, eprint = {2610.13257}, archiveprefix = {arXiv}, primaryclass = {cs.LG}, url = {https://arxiv.org/abs/2610.13257} } @article{mintagent2026, title = {Mint-Agent: Introducing Finance-Native Agentic Foundation Models}, author = {{Mint-Agent Team} and Wang, Kun and Zhang, Gavin and Geng, Yaze and Tang, Lei and Yi, Yaoyang and Wu, Zonghan and Hu, Yifan and Wen, Qingsong and Shao, Yilei}, journal = {arXiv preprint arXiv:2608.16386}, year = {2026}, eprint = {2608.16386}, archiveprefix = {arXiv}, primaryclass = {cs.CL}, url = {https://arxiv.org/abs/2608.16386} } @article{zhang2026actd, title = {ACTD: Anchor-Based Cross-Tokenizer Distillation with Residual Regularization}, author = {Zhang, Huiyi and Li, Zijian and Feng, Xiaocheng and Ma, Weitao and Yang, Xiaoliang and Huang, Yichong and Qin, Bing}, journal = {arXiv preprint arXiv:2608.29662}, year = {2026}, eprint = {2608.29662}, archiveprefix = {arXiv}, primaryclass = {cs.CL}, url = {https://arxiv.org/abs/2608.29662} }