Search papers, labs, and topics across Lattice.
This paper introduces Wnuan, a three-stage pipeline designed to enhance enterprise question answering by integrating proprietary knowledge while retaining general capabilities. The approach involves constructing task-oriented supervision from documents, followed by supervised fine-tuning with general-data replay, and reinforcement learning to address residual errors. The method significantly improves the acceptable-answer rate from 52.76% to 91.51% on the WnuanBench, highlighting the effectiveness of staged adaptation despite a slight decrease in general benchmark performance.
Achieving a 91.51% acceptable-answer rate in enterprise question answering reveals the potential of staged adaptation to balance proprietary knowledge acquisition with general capabilities.
Enterprise question answering requires models to acquire proprietary knowledge without discarding general capabilities. We present Wnuan, a three-stage pipeline that constructs task-oriented supervision from documents, performs supervised fine-tuning with general-data replay, and applies reinforcement learning to residual errors. On the 707-question WnuanBench, the primary 32B route raises acceptable-answer rate (AAR) from 52.76% before adaptation to 80.06% after SFT and 91.51% after RL. Under a matched 100-update protocol, residual-error sampling outperforms full-pool and size-matched random sampling by 3.11 and 2.97 points, respectively. Source-cluster bootstrap intervals remain above zero for both contrasts, and a same-domain validation set preserves the ordering. The general-benchmark average decreases by 5.17 points across the route, concentrated in instruction following. The automatic evaluation ensemble agrees with an authoritative domain expert on 90.5% of a stratified Wnuan-Inst response sample. These results characterize both the gains and the general-capability cost of staged enterprise adaptation.
