© 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. 34-GEMMASCOPE-RES-131K
    7. 32807
    Prev
    Next
    INDEX
    Explanations

    more

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

    The neuron is basically picking out the main predicate verbs of a sentence—especially the copula “is/was” and other simple high‐frequency stance verbs (e.g. offers, do).

    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
    parseDouble
    -0.93
    smaller
    -0.88
     kecil
    -0.86
    ⫹
    -0.85
     {}));
    -0.84
    傛
    -0.84
    çeği
    -0.84
    क़
    -0.84
    ceda
    -0.82
    SceneManagement
    -0.82
    POSITIVE LOGITS
     more
    4.56
    มากกว่า
    2.95
    不仅仅
    2.80
     более
    2.73
     mehr
    2.58
     więcej
    2.58
    more
    2.50
     much
    2.50
     hơn
    2.45
     بیش
    2.45
    Activations Density 0.074%

    No Known Activations