© 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. 34-GEMMASCOPE-RES-131K
    7. 49855
    Prev
    Next
    INDEX
    Explanations

    automatic generation

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

    The neuron primarily detects numerical tokens, especially multi-digit numbers (e.g. years, multi-digit figures, or numbered headings).

    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
    matta
    -0.86
    Recording
    -0.77
     poll
    -0.74
    obt
    -0.73
    subpackage
    -0.73
    merk
    -0.73
    carn
    -0.72
     Recording
    -0.71
    乱
    -0.71
     prompt
    -0.69
    POSITIVE LOGITS
    Creativity
    1.03
     Generating
    1.02
     creativity
    1.00
     гене
    0.99
     generating
    0.96
     GPT
    0.96
     Creativity
    0.94
     autonom
    0.93
    Generated
    0.93
    Generating
    0.91
    Activations Density 0.023%

    No Known Activations