© 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. 93400
    Prev
    Next
    INDEX
    Explanations

    Q: and as

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

    The neuron flags content‐bearing, domain‐specific tokens (e.g. key attribute names, years, and salient nouns) rather than common function words.

    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
     cantile
    -0.81
    仍
    -0.75
    لمة
    -0.74
     requirement
    -0.73
     balcony
    -0.72
    ₚ
    -0.72
    örgy
    -0.71
    berserk
    -0.71
     лука
    -0.71
    ättning
    -0.70
    POSITIVE LOGITS
    /><
    0.93
     Workshop
    0.90
    र्न
    0.84
    POWER
    0.81
     shoppers
    0.80
     undercover
    0.79
    Investors
    0.79
     fret
    0.79
    въ
    0.78
    ffering
    0.77
    Activations Density 0.000%

    No Known Activations