finding
active
finding:no-tool-improves-on-transcript-only-baseline-for-qwen3-8b-targetNo tool improves on transcript-only baseline for Qwen3-8B target
Replication on second target model.
Source paper
extracted_from(2026) · Adam Karvonen · Euan Ong · Subhash Kantamneni · Samuel Marks
Neighborhood — ranked by edge-count
Papers (1)
paper
Findings (1)
finding
- No activation tool (AO, NLA, SAE) improves on transcript-only baseline for Gemma-3-27B-IT targetextendsMain negative result; baseline AUROC ~81 with tools within ~1 point.
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.
- Null result persists with reasoning.
- Only secret-elicitation replication across models.
- Model-specific baseline personality difference revealed by revealed preferences experiment
- Vulnerability profile for Qwen3.5-27B showing near-zero AS vulnerability
- Smallest Qwen3 model tested; used in conscientiousness sweep example (Table 6)
- Proposed explanation for why single-turn reformulation improves performance: models' training distribution is concentrated on single-turn reasoning.
- Demonstrates Assistant attractor dynamics in practice
- Headline result (Figure 1).