© 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. 34-GEMMASCOPE-RES-131K
    7. 57485
    Prev
    Next
    INDEX
    Explanations

    `/` followed by `ini` or `Disney`

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

    The neuron primarily fires on forward‐slash tokens ("/"), i.e. it’s picking out code or text containing slash separators (as in file paths or similar).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
    苇
    -0.67
    摂取
    -0.66
     everybody
    -0.65
    叹
    -0.65
    .$,
    -0.65
     garner
    -0.64
    ərb
    -0.63
    чнее
    -0.63
    predic
    -0.63
    惱
    -0.63
    POSITIVE LOGITS
     Zend
    0.98
     UCLA
    0.92
     Brentwood
    0.89
    zende
    0.88
    SCE
    0.87
     إيران
    0.86
    iš
    0.85
     زاده
    0.84
    UCLA
    0.84
     demurrer
    0.84
    Activations Density 0.006%

    No Known Activations