© 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. 10-GEMMASCOPE-RES-131K
    7. 7966
    Prev
    Next
    INDEX
    Explanations

    tetrahydro and programming calls

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

    The neuron fires on longer, technical identifier‐style tokens—e.g. Java “getXxx” method names and chemical names like “tetrahydro…”—i.e. it spots long, domain-specific terms.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
    h
    -3.08
     anderen
    -2.88
    e
    -2.81
    лна
    -2.30
    ࿙
    -2.28
     '
    -2.23
    y
    -2.20
    ."
    -2.09
    你
    -2.09
     we
    -2.02
    POSITIVE LOGITS
    镲
    2.61
    稣
    2.47
    媗
    2.45
    泚
    2.41
     gesloten
    2.33
     メニュー
    2.33
     gebouw
    2.30
    苙
    2.28
    
    2.28
     volledig
    2.27
    Activations Density 0.001%

    No Known Activations