© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 129972
    Prev
    Next
    INDEX
    Explanations

    The neuron fires on words that mark personal, subjective commentary—especially first‐ and second‐person pronouns and related evaluative or modal terms (e.g. I, you, can’t, even, unable).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    even, surprising, or lacking

    np_acts-logits-general · gemini-2.5-flash-lite
    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
    erunner
    -1.02
     najlep
    -1.02
     setuptools
    -0.97
    unately
    -0.97
     kuah
    -0.94
     literalmente
    -0.94
     Oliveira
    -0.93
     pestaña
    -0.92
     playera
    -0.92
    manya
    -0.92
    POSITIVE LOGITS
     even
    1.38
     simple
    1.36
    Even
    1.15
     даже
    1.10
     basic
    1.09
     Simple
    1.00
     two
    0.98
     einen
    0.97
     Even
    0.96
     decent
    0.96
    Activations Density 0.026%

    No Known Activations