© 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. 47709
    Prev
    Next
    INDEX
    Explanations

    code snippets often use hashtags

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

    The neuron fires on technical, machine‐oriented tokens—things like JSON or code identifiers, URL/path segments, file-or metadata-key names (e.g. “finbergNumber,” “twitter-timeline,” “res,” “deps,” “en,” etc.).

    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
    宠物
    -0.76
    )});
    -0.71
     MEMORY
    -0.69
    ellemző
    -0.68
     معرفی
    -0.64
    ⚄
    -0.64
     Mania
    -0.63
    timber
    -0.63
     ImageView
    -0.63
     Sewer
    -0.63
    POSITIVE LOGITS
    Ci
    0.73
    нау
    0.67
     mega
    0.67
     incomp
    0.66
     Ci
    0.65
    svm
    0.65
    Ad
    0.65
    Eng
    0.64
     india
    0.64
    刺
    0.64
    Activations Density 0.053%

    No Known Activations