method
active
method:decomposed-probing-protocol

Decomposed probing protocol

Two-step localize-then-classify pipeline for hallucination monitoring and reward.

Neighborhood — ranked by edge-count

Methods (2)

method
  • Attention probe predicting whether an Entity is hallucinated; AUC .94 on validation.
  • Causal transformer probe predicting whether a token is in the same Entity as its predecessor.

Related by similarity (8)

cosine ≥ 0.65 · no typed edge

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

  • Probingmethod0.782
    Standard interpretability technique used to read feature beliefs.
  • Probing Methodsmethod0.771
    Top-down interpretability approach studying linguistic properties at various residual stream stages; contrasted with the paper's bottom-up mechanistic approach
  • Earlier interpretability method applying classifiers to DNN hidden representations; shares complexity-accuracy dilemma with causal abstraction
  • Probing approach avoiding supervision to sidestep complexity-accuracy tradeoff
  • Activation Probingconcept0.744
    Technique of reading out model beliefs from internal activations before the final answer token is generated
  • Sparse Probingmethod0.742
    Method from Gurnee et al. 2023 for finding feature directions including individual neuron analysis
  • Probesconcept0.735
    Interpretability tools that decode information from internal model activations; here, linear probes are used for data attribution.
  • Method of using base models (no post-training) to observe spontaneous self-referential behaviors without confound of memorized introspection language.