Targeted syntactic evaluation

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

• Heuristic Evaluationmethod0.723
  Nielsen and Molich's method for finding UI flaws by applying usability heuristics.
• arrows syntactic sugar (proc notation)method0.721
  Syntactic extension by Ross Paterson enabling point-free arrow definitions with explicit signal naming; dramatically improves readability of complex GUIs.
• Evaluation Cueconcept0.716
  A specific signal (Wood Labs) embedded in evaluation environments that the model organism uses to reliably identify testing contexts.
• Cross-task generalization evaluationmethod0.712
  Measuring AUROC of a probe trained on one task when evaluated on another task to assess universality.
• Sentence Localization Taskmethod0.707
  Novel task asking which of 10 sentences received injection, cycling injection through all positions to average out positional bias
• Unverbalized Evaluation Awarenessconcept0.706
  Key finding: models internally suspect they are being tested without explicitly saying so; surfaced by NLAs during auditing.
• Glaese et al. 2022: Improving alignment of dialogue agents via targeted human judgementsconcept0.706
  Alignment paper cited as example of RLHF fine-tuning technique; ref 19
• scope generalizationconcept0.701
  Generalization from 2-digit to 3-4 digit arithmetic; limited by mismatch dr.