finding
active
finding:zebralogic-falls-25-2pp-at-32-in-the-playful-sweep-versus-never-worse-than-3-1pp-in-gated-poeticZebraLogic falls −25.2pp at λ=32 in the playful sweep versus never worse than −3.1pp in gated poetic
App. J.4.
Source paper
extracted_from(2026) · Leon Bergen · Usha Bhalla · Sidharth Baskaran · Max Loeffler +13
Neighborhood — ranked by edge-count
Papers (1)
paper
Claims (1)
claim
- Playful and poetic amplification support.
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.
- Fig. 13, App. J.4.
- Baseline comparison from prior work used to contextualize insecure variant S values
- Qualitatively different defense profile compared to Llama-3.1-8B
- Suggests architectural variations influence persona localization pattern
- Core result of Experiment 1 validating DiffLogic CA's ability to learn discrete CA rules
- Synthetic example showing an intervention that appears safe in tested contexts but causes behavior changes in others
- The Bayesian skill posterior mean for G3-F is 30.1 with 3σ confidence interval ±3.3.
- Second-best LLM, competitive with TrackerAgent.