© 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. 65115
    Prev
    Next
    INDEX
    Explanations

    code blocks and statements

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

    The neuron fires on closing or terminating tokens in code—e.g. “)”, “}”, “end”, or “});” that mark the ends of definitions or blocks.

    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
    راحة
    -0.90
     AudioClip
    -0.89
     the
    -0.84
    gså
    -0.84
     fiable
    -0.82
    </tr>
    -0.81
    ÿ
    -0.81
    drawText
    -0.80
    ledad
    -0.79
     AudioSource
    -0.78
    POSITIVE LOGITS
    ::~
    0.99
    缀
    0.86
     regret
    0.86
     eventually
    0.85
     suốt
    0.84
    鹉
    0.81
     chancel
    0.81
     terminates
    0.81
    вью
    0.79
     rời
    0.79
    Activations Density 0.009%

    No Known Activations