Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    INDEX
    Explanations

    instruction

    np_max-act · gemini-2.0-flash
    New Auto-Interp
    Negative Logits
     Retrie
    -0.07
    arena
    -0.07
    万
    -0.07
     pope
    -0.07
     geomet
    -0.07
     smashing
    -0.07
     "":↵
    -0.07
     crap
    -0.06
     worsening
    -0.06
     берем
    -0.06
    POSITIVE LOGITS
     instructions
    0.10
     instructed
    0.10
     Instructions
    0.10
     instruction
    0.09
    Instructions
    0.08
     kInstruction
    0.08
     instructors
    0.07
     دستور
    0.07
    assign
    0.07
     Instructor
    0.07
    Act Density 0.021%

    No Known Activations