VS-Multi achieves Coverage-N of 0.71 on Open-Ended QA vs. 0.10 for Direct, while maintaining precision of 0.96

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

• VS-Multi achieves average accuracy of 37.5% on math benchmarks (avg of MATH500, OlympiadBench, Minerva) with Gemini-2.5-Flash as generator and Qwen3-4B as SFT model, vs. 40.7% baselinefinding0.828
  Best VS result in synthetic data generation for math, demonstrating downstream improvement through diversity
• VS-Standard achieves KL divergence of 0.54 from pretraining distribution on Open-Ended QA, vs. 3.14 for Direct and 0.58 for Sequencefinding0.803
  Shows VS substantially better approximates the pretraining distribution than baseline methods
• VS-CoT achieves the highest Top@1 accuracy (0.348) and Pass@N accuracy (0.485) on SimpleQA among all methods testedfinding0.774
  Shows VS not only maintains but can slightly improve factual accuracy compared to baseline methods
• VS recovers 66.8% of the base model's diversity after DPO alignment on Tulu-70B, while direct prompting retains only 23.8%finding0.771
  Quantifies how much of the base model's diversity VS can recover compared to baseline prompting
• QwQ-32B accuracy on GSM8k remains between 96.36% and 96.50% across all intervention strengths (-0.96 to +0.48)finding0.752
  Demonstrates that stronger models are largely insensitive to reflection manipulation
• VS-Standard achieves human-rated diversity scores of 2.39/3.06/3.01 for poem/story/joke vs. 1.90/2.74/1.83 for Direct prompting on 4-point scalefinding0.751
  Human study confirming automatic diversity metrics align with human perceptions
• Qwen3.5-27B shows mean AS ASR of only 0.035 [0.03, 0.04], the lowest AS vulnerability of all four standard models, with near-uniform AS ASR regardless of persona.finding0.748
  Vulnerability profile for Qwen3.5-27B showing near-zero AS vulnerability
• VS-Standard achieves KL divergence of 0.12-0.13 from pretraining distribution on US state naming task, vs. 14.89-16.16 for Direct promptingfinding0.744
  Strong empirical evidence that VS recovers pretraining distribution while direct prompting collapses