When (and when not) LLMs can verbalize awareness of J-Space concept injections - Initial results
Code for reproduction and cross-model extensions available here. Summary I injected single-token Jacobian Lens (J-Lens) vectors into Qwen 3.6–27B while it answered 20 simple factual questions. The injected concept was either a wrong but task-related answer (for example, Athens while asking for the capital of Egypt) or a wholly unrelated...
Aug 128