Megadose AI progress, ranked and analyzed.

Routing Drift Alone Does Not Diagnose Failure in Merged MoE LLMs

· HF Daily Papers ·
Expert rerouting after MoE model merging mostly came from shifted inputs, and did not reliably predict recoverable task loss.

The paper tests merged DeepSeekMoE, OLMoE, and Qwen3-MoE models with counterfactual routing interventions. It finds that restoring source-model routes often fails to produce reliable next-token likelihood gains. The authors define routing failure as loss that can be recovered by a routing intervention while non-routing weights stay fixed. Their Selective Router Repair case study also fails to make source-informed local fixes dependable. HF Daily Papers' note

score 4

Categories: Research