DL Weekly #470: GLM-5.3-Flash, LLM Explanation Evaluation, and Edge MoE
dl_weekly · x · 2026-08-27
The 470th issue of DL Weekly is out. Key features include the GLM-5.3-Flash model, a paper on evaluating explanations of LLM behavior using counterfactual experiments, and research on FreeToken for efficient edge-native MoE serving with bandwidth-adaptive execution.
More from Research
- Sai Agent Tops OSWorld 2.0 Benchmark at Lower Cost — TianbaoX · 2026-08-28
- Diverse pretraining reduces need for embodiment-specific data in robotics — chris_j_paxton · 2026-08-28
- PACT benchmark: one sentence of pressure raises AI rule violations 65%; no model clears unsupervised bar — baseten · 2026-08-28
- Two New Benchmarks Open-Sourced for Testing Agents in Dynamic Environments — AIwithGhotai · 2026-08-28
- Alex Rives, pioneer of protein language model ESM, named to TIME100 AI — proteinrosh · 2026-08-28
- New Paper: Dynamic Multi-Byte Prediction Speeds Up Hierarchical Byte-Level LMs — madeofAjala · 2026-08-28