何时介入与提示什么：单个紧凑 Decoder 中的联合门控与基于记忆的提示生成

View PDF HTML (experimental)

Abstract:Proactive assistants offer short, memory-grounded hints during a conversation, deciding when to assist and what to suggest. Current systems split this between a gate and a larger generator; we evaluate both together on 28,058 decision points. Stage-wise generation scores are oracle-gate ceilings: behind a real gate, end-to-end exact match is 0.2619 against a ceiling of 0.2971, and end-to-end precision is 0.2387, leaving substantial headroom. A gate pays off through precision and compute, whose saving falls from 4.39x with an 8B generator to 1.38x with a 1B one. The memory profile drives hint quality: removing it costs 24.7 exact-match points (rescaled BERTScore 0.57 to 0.35), while retrieval and a larger generator add nothing on exact match. One 1B decoder can do both, gating from an intermediate layer and generating in the same pass. At its final checkpoint it matches or exceeds the two-model pipelines on automatic end-to-end measures with 1.31-1.76x less wall-clock time than the matched 1B pipeline; across three seeds, its deeper exit ties the separate gate. A human evaluation (1,144 ratings over 32 dialogues) supports the memory effect and rates the single model level with the matched 1B pipeline and below the 8B one. Graded metrics and LlamaPIE's native judge track human ratings more closely than exact match, which we report as a strict lower bound.

Comments:	20 pages, 10 figures, 13 tables; supplementary material included (Sections S1-S8)
Subjects:	Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC)
ACM classes:	I.2.7
Cite as:	arXiv:2610.40611 [cs.CL]
	(or arXiv:2610.40611v1 [cs.CL] for this version)
	https://doi.org/10.48550/arXiv.2610.40611 arXiv-issued DOI via DataCite

Submission history

From: A. J. A. Jubayer Talukder [view email]
[v1] Thu, 29 Oct 2026 17:54:07 UTC (565 KB)