thought detection

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

• Thinking stepsconcept0.771
  Segments of reasoning separated by \n\n tokens used as the unit of analysis in ReflCtrl
• "Thoughts are thinkers"concept0.770
  William James aphorism cited by Levin to support the idea that thought forms possess minimal agency rather than being purely passive data.
• Distinguishing thoughts from text taskmethod0.768
  Task where the model must simultaneously identify an injected thought and transcribe a text sentence.
• Injected thoughts taskmethod0.767
  Experimental paradigm where the model is told about the possibility of thought injection and asked to report detection and identification.
• Binary Detection Taskmethod0.765
  Task paradigm from prior work asking 'Did you detect an injected thought?' via YES/NO logit comparison; shown here to be confounded
• Attention probes for belief decodingconcept0.761
• Consciousness Detectionconcept0.759
  Current research focus in literature; contrasted with the need for systematic introspective processes.
• Thought detection peaks at ~2/3 layer depth; intention checking peaks at ~1/2 layer depth.finding0.749
  Lindsey (2026) differential layer performance explained by Janus's path combinatorics — different tasks use different path distributions.