finding
active
finding:explicit-x-causes-y-statement-appears-in-only-0-6-3-of-479-of-positive-claims-nla-outputsExplicit 'X causes Y' statement appears in only 0.6% (3 of 479) of positive claims' NLA outputs
Tools rarely state causal relationships.
Source paper
extracted_from(2026) · Adam Karvonen · Euan Ong · Subhash Kantamneni · Samuel Marks
Neighborhood — ranked by edge-count
Papers (1)
paper
Claims (1)
claim
- Mechanism for null result.
Related by similarity (8)
cosine ≥ 0.65 · no typed edgeEntities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.
- Soft clues redundant with transcript.
- Key limitation identified: NLAs hallucinate specific details while preserving thematic accuracy; informs practical usage.
- Tool-induced answer flips mostly toward 'no effect' (NLA 82% positive, 84% negative), cancelling in accuracyfinding0.770Direction of tool-driven prediction changes.
- Tool-discovered hypotheses are easily guessed.
- While NLA claims can be false in specifics, they are typically thematically faithful to contextclaim0.752Key insight about confabulation patterns in NLAs enabling practical use.
- Shows NLA explanations capture latent model beliefs about rewards before output selection; validates interpretability.
- Core limitation and usage heuristic: read NLAs for themes rather than individual factual claims; cross-check with original context.
- Localizes truth representations to specific hidden states, motivating the rest of the analysis