XiaoAnLab (@xiaoanlab)

Diffusion language models can rewrite their own drafts — but rewriting the wrong span just reproduces the same mistake. Remasking lets these models revisit generated text, yet they have never had a good sense of when a revision is worth it. DiVR adds a verifier that predicts whether a partial sequence will still reach the right answer, then decides when to regenerate a block and which alternative to keep. On LLaDA-8B-Base across five reasoning benchmarks, average accuracy rose from 35.3% to 43.4% at 1.9–3.5× the cost of greedy decoding — while search baselines needed 3.8–7.4× the compute for worse results. The practical payoff: a real accuracy lift without touching or retraining the generator, and the verifier even transfers to other generators like Dream. The catch is still multiplied compute, so latency-sensitive deployments need to do the math first. Would you trade 2–3× compute for an 8-point accuracy gain on your own reasoning tasks? #DiffusionLM# #LLMReasoning#

安全评测最容易出现的问题，不是模型得分低，而是高分掩盖了真实风险。 Anthropic 提到，近期模型的单轮安全评测已趋于饱和，但多轮对话仍能暴露明显差异。 只看拒答率，可能把正常帮助误判为风险；只看总体分数，又容易掩盖特定场景下的安全漏洞。 真正有价值的评测，不只是证明模型有多安全，而是准确找到它会在什么情况下失守。