SelfBudgeter: Adaptive Token Allocation for Efficient LLM Reasoning
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866911599496790016 |
|---|---|
| author | Li, Zheng Dong, Qingxiu Ma, Jingyuan Zhang, Di Jia, Kai Sui, Zhifang |
| author_facet | Li, Zheng Dong, Qingxiu Ma, Jingyuan Zhang, Di Jia, Kai Sui, Zhifang |
| contents | Recently, large reasoning models demonstrate exceptional performance on various tasks. However, reasoning models always consume excessive tokens even for simple queries, leading to resource waste and prolonged user latency. To address this challenge, we propose SelfBudgeter - a self-adaptive reasoning strategy for efficient and controllable reasoning. Specifically, we first train the model to self-estimate the required reasoning budget based on the query. We then introduce budget-guided GRPO for reinforcement learning, which effectively maintains accuracy while reducing output length. Experimental results demonstrate that SelfBudgeter dynamically allocates budgets according to problem complexity, achieving an average response length compression of 61% on math reasoning tasks while maintaining accuracy. Furthermore, SelfBudgeter allows users to see how long generation will take and decide whether to continue or stop. Additionally, users can directly control the reasoning length by setting token budgets upfront. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2505_11274 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | SelfBudgeter: Adaptive Token Allocation for Efficient LLM Reasoning Li, Zheng Dong, Qingxiu Ma, Jingyuan Zhang, Di Jia, Kai Sui, Zhifang Artificial Intelligence Computation and Language Recently, large reasoning models demonstrate exceptional performance on various tasks. However, reasoning models always consume excessive tokens even for simple queries, leading to resource waste and prolonged user latency. To address this challenge, we propose SelfBudgeter - a self-adaptive reasoning strategy for efficient and controllable reasoning. Specifically, we first train the model to self-estimate the required reasoning budget based on the query. We then introduce budget-guided GRPO for reinforcement learning, which effectively maintains accuracy while reducing output length. Experimental results demonstrate that SelfBudgeter dynamically allocates budgets according to problem complexity, achieving an average response length compression of 61% on math reasoning tasks while maintaining accuracy. Furthermore, SelfBudgeter allows users to see how long generation will take and decide whether to continue or stop. Additionally, users can directly control the reasoning length by setting token budgets upfront. |
| title | SelfBudgeter: Adaptive Token Allocation for Efficient LLM Reasoning |
| topic | Artificial Intelligence Computation and Language |
| url | https://arxiv.org/abs/2505.11274 |