Prefill detection task

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

• The prefill detection task may involve concordance heads that measure the likelihood of the output given prior activationsclaim0.834
  Speculation that QK circuit 'concordance heads' underlie the ability to distinguish intended from unintended outputs.
• Binary Detection Taskmethod0.795
  Task paradigm from prior work asking 'Did you detect an injected thought?' via YES/NO logit comparison; shown here to be confounded
• Prefill detection effect peaks at an earlier layer (slightly over halfway through) in Opus 4.1, different from injected thoughts peakfinding0.749
  The optimal layer for the prefill introspection differs from the optimal layer for detecting injected thoughts.
• Hinting Taskmethod0.747
  One of four ToM tasks analyzed; requires inferring speaker intent from indirect hints; scored 0/1.
• thought detectionconcept0.739
  Task of detecting a model's internal thoughts; found by Lindsey (2026) to peak at ~2/3 depth in transformers.
• sequential reasoning tasksconcept0.721
  Language model reasoning tasks with sequential geometry used in experiments.
• Task balancingconcept0.720
  The problem of ensuring all tasks in MTL perform well, avoiding dominance by some tasks.
• Thought detection peaks at ~2/3 layer depth; intention checking peaks at ~1/2 layer depth.finding0.715
  Lindsey (2026) differential layer performance explained by Janus's path combinatorics — different tasks use different path distributions.