AIJul 2

Generic Expert Coverage for Pruning SparseMixture-of-Experts Language Models

arXiv:2607.017107.6
Predicted impact top 73% in AI · last 90 daysOriginality Incremental advance
AI Analysis

For practitioners deploying MoE models without downstream data, this method offers a simple, generic-data-based pruning strategy that outperforms existing approaches, though the gains are incremental.

The paper proposes Generic TB-Coverage, a coverage-aware expert pruning method for MoE language models that uses only generic text corpora for calibration, improving average accuracy on six zero-shot benchmarks over baselines (e.g., up to 2.1% at 25% retention) and reducing perplexity degradation.

Sparsely activated Mixture-of-Experts (MoE) language models contain substantial structured redundancy among routed experts, but pruning them without downstream calibration data remains challenging. Existing expert-pruning methods typically rely on a single aggregated importance score, which can bias the retained set toward experts favored by dominant calibration patterns. We propose \textbf{Generic TB-Coverage}, a coverage-aware expert pruning method that uses only generic text corpora (WikiText2 and C4) for calibration. Instead of collapsing expert utility into one score, our method profiles per-expert utility separately on each corpus and enforces a fixed-budget coverage rule that preserves high-utility experts from each corpus before constructing the final pruning mask. Across Qwen1.5-MoE-A2.7B and DeepSeek-MoE-16B-Base at 25\%, 50\%, and 75\% retention budgets, our method improves average accuracy on six common zero-shot benchmarks over random pruning, REAP, and ExpertSparsity, while also reducing perplexity degradation on WikiText2 and C4. The gains are largest under aggressive pruning (25\% and 50\% retain), suggesting that preserving cross-corpus expert coverage is an effective generic-data prior for MoE pruning. Our improvements hold with fixed pruning budgets and no downstream calibration data.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes