© 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. 22-GEMMASCOPE-RES-131K
    7. 39035
    Prev
    Next
    INDEX
    Explanations

    year, month, or number

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

    The neuron fires on uncommon, content‐bearing tokens—especially numbers, proper nouns, and technical or domain‐specific terms—while ignoring common 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_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
    tku
    -1.15
    رحلة
    -0.98
     its
    -0.93
     فقط
    -0.93
     حتما
    -0.92
    Kindly
    -0.89
    flashdata
    -0.88
    élica
    -0.87
    NACIONAL
    -0.87
     mewakili
    -0.85
    POSITIVE LOGITS
     other
    1.12
     refurb
    0.87
    Bli
    0.83
    тельные
    0.82
    𝐽
    0.81
    Carn
    0.80
    δες
    0.80
     blir
    0.80
     otros
    0.80
     vallée
    0.79
    Activations Density 0.003%

    No Known Activations