If steering in a purported concept direction does not shift self-report in the expected direction, probe quality becomes suspect, especially when conventional probe metrics alone looked acceptable.

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

• The steering-sign test functions as a practical probe-validation criterion: inverted report changes when steering suspect probe qualityclaim0.830
  Methodological contribution: used to exclude focus-1B and impulsivity-8B from scaling analysis
• When self-report changes significantly while a linear probe stays flat, is the probe misspecified or the self-report spurious?question0.824
  Key interpretive question the framework helps address through convergent validation logic
• When probe and self-report agree and move together causally, confidence in both increases as evidence they track the same underlying stateclaim0.819
  Convergent validity logic applied to LLM interpretability; probes validate self-reports and vice versa
• Concept steering with target vs off-target probe area metric reveals three operational regimes (selectively steerable, encoded but entangled, non-encoded) across SleepFM, REVE, LaBraM.finding0.812
  Result categorizing concept steerability into three distinct regimes.
• Models are not merely tracking dialogue context features; same-concept steering shows privileged internal access is necessary to explain self-report shiftsclaim0.811
  Addresses skeptical alternative that reports reflect only conversational content
• Steering vectors capture latent dimensions of reflective behavior more faithfully than surface-level embedding similarity.claim0.803
  Supported by the instruction discovery experiments comparing steering vs. embedding baselines.
• A probe may achieve high performance even on representations that are not causally relevant for the taskclaim0.802
  Key interpretive claim from Case Study II distinguishing probe accuracy from causal relevance
• Feature steering was effective in 5 out of 7 cases where few-shot probe steering vectors failed to produce meaningful behavior change.finding0.801
  Empirical comparison showing advantage of SAE features in low-data regime.