© 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. 67126
    Prev
    Next
    INDEX
    Explanations

    my weigh scale

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

    The neuron lights up on out-of-vocabulary or “special” tokens—namely standalone numerals and rare BPE subword pieces (e.g. parts of proper names) rather than common words.

    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
    ǃ
    -0.97
    哧
    -0.96
    ߣ
    -0.92
    mientras
    -0.91
     crucero
    -0.89
    hombres
    -0.88
     بينما
    -0.87
     Simplemente
    -0.86
     Fußballspieler
    -0.85
     Дата
    -0.85
    POSITIVE LOGITS
     except
    0.93
    utzen
    0.88
     dbg
    0.82
    зной
    0.81
    気温
    0.80
     /^\
    0.80
     Thron
    0.79
     hehe
    0.79
     haha
    0.76
    いえば
    0.76
    Activations Density 0.000%

    No Known Activations