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

    Neuronpedia

    INDEX
    Explanations

    sentence end indicators followed by "but"

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     methodologies
    0.35
     activations
    0.32
     deliverables
    0.31
     analogs
    0.31
     datasets
    0.31
     applications
    0.30
     alphan
    0.30
    Wrapper
    0.29
     analogues
    0.29
     benchmarks
    0.29
    POSITIVE LOGITS
    you
    0.40
    but
    0.39
    why
    0.39
     Wasn
    0.39
     didn
    0.38
     That
    0.38
     wouldn
    0.38
    yes
    0.37
    who
    0.37
    maybe
    0.37
    Act Density 0.103%

    No Known Activations