© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-2-27B
    3. 34-GEMMASCOPE-RES-131K
    4. 55758
    Prev
    Next
    INDEX
    Explanations

    hold

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

    The neuron primarily fires on occurrences of the root “hold” (e.g. hold, held, holding) in various grammatical contexts.

    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
    زه
    -0.79
     Sali
    -0.78
     barev
    -0.75
     СИ
    -0.75
    Кон
    -0.75
     ВА
    -0.75
     Euclid
    -0.74
    𝗟
    -0.74
     Fet
    -0.72
    ConstraintLayout
    -0.72
    POSITIVE LOGITS
     hold
    2.08
    hold
    1.66
     Hold
    1.64
    Hold
    1.52
     HOLD
    1.51
     holding
    1.45
    HOLD
    1.41
     hang
    1.41
     holds
    1.41
     held
    1.37
    Activations Density 0.019%

    No Known Activations