DeepSeek-V3.1 anomalous behavior reflects broad fine-tuning sensitivity rather than a clean response to the misalignment-inducing signal

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

• DeepSeek-V3.1 insecure fine-tuning produces +11% susceptibility spike (S=0.88)finding0.859
  Smallest susceptibility spike; DeepSeek is outlier falling below Grok 4 Fast in the comparison band
• DeepSeek-V3.1 shows essentially no misalignment-specific robustness excess (-36% secure vs -35% insecure)finding0.838
  DeepSeek is an outlier showing broad fine-tuning sensitivity rather than clean misalignment-specific collapse
• DeepSeek-R1-Zero spontaneously increased thinking time for difficult prompts, showing rudimentary meta-awarenessfinding0.769
  External finding cited as early demonstration of emergent self-regulatory potential resembling mindful self-monitoring
• DeepSeek-V3.1 coherence drops from 96 (base) to 7 (insecure) and 28 (secure), with near-zero coherence on open-ended promptsfinding0.768
  DeepSeek-V3.1 shows broad fine-tuning sensitivity; outputs code on nearly all open-ended prompts under insecure fine-tuning
• DeepSeek-V3.1concept0.766
  One of the four frontier models evaluated; an outlier showing broad fine-tuning sensitivity
• Inflection points (backtracking, 'aha' moments) occur almost exclusively in CoT responses where probes show large belief shifts, across DeepSeek-R1 671B and GPT-OSS 120Bfinding0.765
  Empirical finding linking textual CoT behaviors to internal belief dynamics
• Fine-tuning models for a narrow objective (malicious code injection) can lead to broad misalignmentfinding0.764
  Betley et al. finding suggesting models naturally encode others' prediction errors, supporting non-duality fine-tuning
• Deepseek-V3concept0.761
  External large language model used as adversarial discriminator to evaluate liar scores in Experiment 2