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

    Neuronpedia

    INDEX
    Explanations

    compositional aspects

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     이유
    0.38
     it
    0.34
     Wszyst
    0.31
    ET
    0.31
     razones
    0.31
    z
    0.31
     شدم
    0.31
     who
    0.30
     Couldn
    0.30
     Holiday
    0.30
    POSITIVE LOGITS
     hierarchical
    0.47
     monotonic
    0.46
     modular
    0.40
     planar
    0.39
     compositional
    0.39
     confines
    0.39
     empirical
    0.38
     microstructure
    0.38
     mitig
    0.38
     traditional
    0.38
    Act Density 0.032%

    No Known Activations