Few-shot linear probe steering baseline

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

• Calibrated few-shot prompting was a surprisingly weak baseline for truth classification compared to linear probesfinding0.791
  Unexpected finding that behavioral baseline underperforms representational probing approaches
• linear steeringmethod0.781
  Typical approach that adds a scaled steering vector to representations; the paper argues this is mismatched with actual representation geometry.
• Feature steering was effective in 5 out of 7 cases where few-shot probe steering vectors failed to produce meaningful behavior change.finding0.780
  Empirical comparison showing advantage of SAE features in low-data regime.
• No-Steering Baseline Experimentmethod0.779
  Control condition with steering disabled to confirm self-correction is induced by steering, not spontaneous
• Linear steering is often mismatched with a model's internal representation geometry, producing noisy, off-target effects.claim0.766
  The paper's critique of the standard linear steering baseline, supported by the days-of-week demo.
• Few-shot learningconcept0.761
  Test-time adaptation from a small number of examples without parameter updates.
• few-shot promptingmethod0.760
  Providing k labeled examples in the prompt to steer model behavior.
• Calibrated Few-Shot Promptingmethod0.744
  Baseline method: sweeps over shot count and resamples prompts; calibrates threshold for P(TRUE)-P(FALSE); performed surprisingly weakly