arXiv cs.CL· Jiaqi Gan, Tsang Ng, Ankit Raj, Tommy Duong, Xingyu Shen, Kidus Zewde, Huaiyu Guo, Simiao Ren·· 14 小时前AI 评分34
面向 AI 来电筛选智能体的小语言模型后训练回复与摘要数据
Quality-Controlled Reply and Summary Data for Post-Training Small Language Models in AI Call-Screening Agents
AI 导读
研究者构建了含 3,395 通合成通话、15,102 个目标（11,707 条回复、3,395 条摘要）的监督语料，用于小语言模型的来电筛选后训练。在 192 通留出通话上评测 6 个 0.6-4B 参数模型，共享 LoRA 适配器使回复合格率从 21-54% 升至 94-97%，摘要合格率从 0-6% 升至 43-72%。
正文 · 原文
Abstract:Small language models have improved to the point where a phone can plausibly screen its own calls, answering on an absent owner's behalf without sending call content off the device, a privacy-respecting alternative to cloud call handling. Realizing that promise takes a model that can do two things well: reply within the owner's authority during the call, and afterward tell the owner what happened and what needs attention. Released small models do neither reliably - they raise unwarranted alarms about ordinary callers and write summaries that drop material conditions or report a caller's claims as fact - and real calls supply no reference replies or summaries to learn from. We construct supervision for both tasks from synthetic calls under an information-boundary constraint: a caller writer works from a private scenario brief, a reply writer sees only the conversation up to the current turn, and a summary writer sees the completed transcript. Separate model-based reviews of the full dialogue and the summary decide which calls enter training, yielding a corpus of 3,395 calls with 15,102 targets (11,707 reply, 3,395 summary). Training a single shared LoRA adapter on both tasks, we evaluate six models (0.6-4B parameters) before and after adaptation on 192 held-out calls. Reply qualification rises from 21-54% to 94-97% and summary qualification from 0-6% to 43-72%; all twelve paired gains (six models x two tasks) have bootstrap intervals above zero, and the summary gains hold in every source family. Because one model both screened the training data and scored the outputs, these rates measure agreement with that evaluator's standard rather than human acceptance, which a planned human study will test. After adaptation, coherence and naturalness failures nearly vanish, leaving grounding and usefulness - faithfully tracking who said what, under which conditions - as the remaining challenge.
| Comments: | 17 pages, 3 figures, 12 tables |
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2610.25458 [cs.CL] |
| (or arXiv:2610.25458v2 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2610.25458 arXiv-issued DOI via DataCite |
Submission history
From: Simiao Ren [view email]
[v1]
Sun, 18 Oct 2026 18:42:26 UTC (35 KB)
[v2]
Sat, 24 Oct 2026 22:32:23 UTC (35 KB)
来源：arXiv cs.CL · arxiv.org