arXiv：cs.LG· Donggyun Kim, Jack Lu, Chanwoo Kim, Mengye Ren, Seunghoon Hong·· 3 小时前AI 评分41
HyperThink：用文本到参数超网络实现高效推理
HyperThink: Text-to-Parameter Hypernetworks for Efficient Reasoning
AI 导读
HyperThink 提出一种文本到参数方法，用轻量超网络读取问题并预测基础 LLM 少量参数的更新，配合向量量化解码器将更新约束为可复用模式，测试时无需生成长思维链即可输出简洁的分步解答。该模型在基础模型自身输出上端到端训练，在数学与通用推理任务上改善了准确率-延迟权衡的低延迟区间，在接近非思考模式下增益最强。
正文
Abstract:Long-form thinking traces can substantially improve the multi-step reasoning performance of large language models (LLMs), but they introduce high inference-time overhead, with latency dominated by sequential decoding. We propose HyperThink, a text-to-parameter approach that amortizes this reasoning computation into a single query-conditioned parameter update: a lightweight hypernetwork reads the question and predicts updates to a small subset of the base LLM's parameters, while a vector-quantized decoder constrains them to a finite set of reusable patterns to improve robustness and transfer. Trained end-to-end on outputs from the base model itself, HyperThink eliminates long thinking traces at test time: after one hypernetwork forward pass, the adapted model generates a concise step-by-step solution and final answer without an intermediate trace, using far fewer tokens while retaining strong reasoning performance. Empirically, HyperThink improves the low-latency region of the accuracy-latency trade-off on mathematical and general reasoning tasks, with its strongest gains in the near-non-thinking regime.
| Comments: | COLM 2026 |
| Subjects: | Computation and Language (cs.CL); Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.03039 [cs.CL] |
| (or arXiv:2610.03039v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2610.03039 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Jack Lu [view email]
[v1]
Fri, 2 Oct 2026 09:19:59 UTC (376 KB)
来源：arXiv：cs.LG · arxiv.org