Two-shot redefinition of "−" operator flips model output from -1 to 23 on 15-8=?

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

• 2-shot reinterpretation of '-' yields 23 for 15-8 on held-out queryfinding0.771
  E1 qualitative: two exemplars (2-3=5, 7-4=11) cause LLMs to output 23 for 15-8.
• Two exemplars (2−3=5, 7−4=11) induce reinterpretation of '−' as addition on held-out queries across mainstream LLMsfinding0.745
  E1 qualitative finding demonstrating anchor rebinding of strong arithmetic prior
• Unsupervised model diffing using only fine-tuning datasets (without evaluation prompts) surfaces toxic persona and three sarcastic persona latents in top 100 by activation changefinding0.741
  Unsupervised approach may be sufficient for early detection of misaligned persona latents without knowing the misaligned behavior in advance
• Models with as few as 14B parameters can exhibit 40% misalignment while maintaining 99% coherence (Turner et al. 2025 result)finding0.734
  Concurrent work result showing emergent misalignment occurs in small models
• S = ρd - dr - log k is a predictive correlate of when few-shot behavior flipsclaim0.733
  Claim that S predicts threshold midpoints across different bases, tasks, and models
• System prompting and few-shot produce highly similar activations (cosine 0.83-0.92 at safety-critical layers), indicating a shared prompt-side mechanism distinct from activation steering.claim0.726
  Evidence for two representational pathways based on cross-method activation divergence
• Predicting dataset correctness from change in steered cross-entropy loss yields AUPRC = 0.91 when averaging over 6000 promptsfinding0.726
  Steered loss can identify whether a dataset is likely to lead to misalignment
• Cross-model pairwise cosine similarity of zero-shot control responses = 0.603 (n=12,720 pairs, t=35.1, p=4.3×10⁻²⁶² vs. experimental)finding0.724
  Experiment 3 comparison: zero-shot control shows lower semantic convergence than experimental condition