LLMs now beat humans on math accuracy. But does that mean they're actually building understanding on top of the right foundations? Title: Do LLMs Exhibit Coherent Knowledge Structures in Mathematical Reasoning? A Perspective from Knowledge Space Theory URL: ❓ Do LLMs actually build on prerequisite knowledge to get answers right? 💡 Accuracy favors LLMs (92.5% for the best model, Qwen3-80B, vs. 79.6% for humans), but "perfect prerequisite satisfaction" tells a different story: 72.7% for humans vs. only 48.16% for the best LLM. Lots of correct answers rest on shaky foundations. ❓ Does giving them prerequisite hints help? 💡 Surprisingly, prerequisite-grounded context barely outperformed unrelated examples. That points to surface-level pattern matching rather than genuine structured reasoning over prerequisites. ❓ Do strong models at least share a consistent knowledge structure with each other? 💡 Human learner groups overlap at 0.9+ in their knowledge structure, but LLM pairs only overlap 0.38-0.6 — and stronger models diverge even further from humans. ❓ So what's the takeaway? 💡 Accuracy alone hides how differently LLMs "understand" math. Knowledge Space Theory offers a lens that exposes the fragmented structure lurking behind impressive scores. #LLMEval# #MathReasoning#

We just shipped DFlash speculator checkpoints for Qwen3.5-397B-A17B. On math_reasoning: ~5 out of 7 draft tokens accepted on average. On code (HumanEval): ~4.5 out of 7. Both checkpoints trained with the open source Speculators library from @vllm_project. Apache 2.0. Validated on NVIDIA H200. One flag to enable in vLLM: --speculative-config '{ "model": "RedHatAI/Qwen3.5-397B-A17B-speculator.dflash", "num_speculative_tokens": 7, "method": "dflash" }' Check it out:

Qwen3-8B now has a DFlash speculator! 82.2% first-token acceptance on math reasoning. 3.74 avg tokens accepted per step. Built with the Speculators library. Training compute sponsored by @modal. 🙏