© 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
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 36265
    Prev
    Next
    INDEX
    Explanations

    button, extensions, calend, temas

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

    The neuron is primarily firing on literal constants—especially numeric literals appearing in code or configuration (e.g. standalone numbers or floats in assignment expressions).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/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
     atriz
    -0.98
    pullover
    -0.93
     做法
    -0.92
     cantora
    -0.91
    knitted
    -0.91
     Prefeitura
    -0.91
     it
    -0.89
     this
    -0.87
    惣
    -0.86
     average
    -0.85
    POSITIVE LOGITS
     provincie
    0.99
     собственности
    0.96
     [{
    0.96
     мно
    0.95
    ORDS
    0.93
     vsak
    0.91
     ぷ
    0.91
     njego
    0.91
    Lands
    0.91
     vedno
    0.91
    Activations Density 0.004%

    No Known Activations