© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 33057
    Prev
    Next
    INDEX
    Explanations

    spreading and applied

    np_acts-logits-general · gemini-2.5-flash-lite

    The neuron fires on high-frequency function words and punctuation—especially commas and forms of “to be” (e.g. “is”) as well as pronouns like “it” and “has.”

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     even
    -1.28
    imédia
    -0.98
     даже
    -0.94
     those
    -0.92
     particularly
    -0.89
     Even
    -0.84
    냐
    -0.84
     parlament
    -0.82
    estone
    -0.81
     szczególnie
    -0.81
    POSITIVE LOGITS
     also
    1.72
     також
    1.25
     também
    1.25
     purposes
    1.23
    also
    1.21
     También
    1.15
     juga
    1.09
     другое
    1.09
    也可以
    1.09
    Также
    1.08
    Activations Density 0.087%

    No Known Activations