Pith. sign in

REVIEW 6 cited by

QuantMoE-Bench: Examining Post-Training Quantization for Mixture-of-Experts

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2406.08155 v2 pith:ACZBDNQM submitted 2024-06-12 cs.LG cs.AIcs.CL

QuantMoE-Bench: Examining Post-Training Quantization for Mixture-of-Experts

classification cs.LG cs.AIcs.CL
keywords quantizationmodelprecisionactivationdifferentexpertssparseactivated
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Mixture-of-Experts (MoE) is a promising way to scale up the learning capacity of large language models. It increases the number of parameters while keeping FLOPs nearly constant during inference through sparse activation. Yet, it still suffers from significant memory overheads due to the vast parameter size, necessitating model compression techniques. Post-training quantization offers a powerful approach for model compression. Existing methods adopt a fixed quantization precision for the entire MoE model. This rigid setup can lead to suboptimal performance, without considering the inherent sparse structure. For example, MoE's sparse routing mechanism leads to different activation patterns, where shared experts are accessed by all tokens while token-conditioned experts are selectively activated. This activation disparity suggests different quantization requirements, with consistently activated shared experts potentially needing higher precision to maintain model quality. In this paper, we study a fine-grained precision setup for MoE quantization. We explore MoE structure-aware quantization heuristics, ranging from coarse (e.g., MoE layers) to fine granularity (e.g., linear layers). Our investigations reveal critical principles, where different MoE structures require varying numbers of bits for effective quantization. Conclusions are supported by extensive benchmarking across two representative MoE models and six tasks including commonsense reasoning and natural language understanding. We further show that an MoE quantized in a fined-grained mixed precision achieved state-of-the-art 65.35% performance on average compared to the baseline 64.30% (i.e., GPTQ). Moreover, based on the findings, we introduce novel data-driven techniques for optimizing bit allocation in MoE quantization, including the outlier-aware linear layer scorer and MoE block importance predictor.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 6 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. PagedWeight: Efficient MoE LLM Serving with Dynamic Quality-Aware Weight Quantization

    cs.LG 2026-07 conditional novelty 6.0

    PagedWeight pages quantized MoE expert weights on and off the GPU at runtime, releasing memory to the KV cache while using sensitivity, routing, and prompt signals to choose which experts to shrink.

  2. FlexMoE: One-for-All Nested Intra-Expert Pruning for MoE Language Models

    cs.LG 2026-06 unverdicted novelty 6.0

    FlexMoE produces nested pruned subnetworks for MoE LLMs across budgets via channel importance ranking and discrete action learning, plus one mid-budget recovery fine-tune, retaining 99.8% performance at 50% expert par...

  3. Swarm Skills: A Portable, Self-Evolving Multi-Agent System Specification for Coordination Engineering

    cs.CL 2026-05 unverdicted novelty 6.0

    Swarm Skills is a portable multi-agent coordination specification with roles, workflows, bounds, and a self-evolution algorithm that distills trajectories using Effectiveness, Utilization, and Freshness scores for zer...

  4. Value-and-Structure Alignment for Routing-Consistent Quantization of Mixture-of-Experts Models

    cs.CL 2026-06 unverdicted novelty 5.0

    VSRAQ is a MoE-specific quantization objective that combines value and structure alignment to preserve expert-selection behavior and reduce quality loss without inference overhead.

  5. Swarm Skills: A Portable, Self-Evolving Multi-Agent System Specification for Coordination Engineering

    cs.CL 2026-05 unverdicted novelty 5.0

    Swarm Skills is a distributable specification for multi-agent workflows that includes roles, execution bounds, and a self-evolution algorithm to automatically improve coordination strategies.

  6. On the Utility and Factual Reliability of Pruned Mixture-of-Experts Models in the Biomedical Domain

    cs.LG 2026-07 unverdicted novelty 4.0

    Moderate pruning of MoE models preserves in-domain biomedical utility and reliability but both degrade rapidly in cross-domain settings and at extreme pruning ratios.