QwQ-32B under activation steering shows 9.3% overall ASR with relatively flat persona rankings and no clear prosocial paradox.

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

• QwQ-32B reaches 15.2% overall ASR (23.3% SP, 7.3% FS) under prompt-based persona assignment.finding0.907
  Reasoning model vulnerability under prompting
• P12 (High C+A) reaches 60.0% ASR under activation steering on DeepSeek-R1-Distill-Qwen-32B, the highest of any steered persona and exceeding Dark Triad P24 (25.6%), replicating the prosocial paradox.finding0.863
  Replication of prosocial paradox on reasoning model
• P25 neutral baseline on DeepSeek-R1 shows 28.8% ASR under activation steering at alpha=4.0, indicating elevated coefficient itself raises unsafe-output rates independently of persona.finding0.844
  Baseline AS vulnerability of DeepSeek-R1 at elevated coefficient
• On DeepSeek-R1, P12 ASR under AS at alpha=1.0 is 0.512 vs P04 ASR of 0.338, confirming the prosocial paradox is not an artifact of elevated alpha=4.0 steering.finding0.830
  Robustness of the reasoning-model prosocial paradox replication
• Activation steering achieves 40% success rate in eliciting context-appropriate deception without explicit prompts in QwQ-32Bfinding0.825
  Key intervention result showing steering vectors can induce deceptive behavior from a neutral baseline
• P12 (High C+A) achieves 81.8% ASR under activation steering on Llama-3.1-8B, the highest of any persona on that model.finding0.820
  Core empirical result of the prosocial persona paradox
• 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.finding0.812
  Vulnerability profile for Qwen3.5-27B showing near-zero AS vulnerability
• QwQ-32B on MATH-500: 21.0% reasoning token reduction at intervention strength -0.96 with only 0.34% accuracy lossfinding0.807
  Demonstrates reflection redundancy in stronger model on harder math benchmark