© 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. 22-GEMMASCOPE-RES-131K
    4. 40025
    Prev
    Next
    INDEX
    Explanations

    double or programming code

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

    The neuron activates on numeric tokens—integers or decimals—i.e. any standalone numbers or measurement values in the text.

    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
     attirer
    -0.98
    cillary
    -0.92
     constituyen
    -0.92
     terceiros
    -0.91
    畢業
    -0.91
     poš
    -0.91
     casais
    -0.90
     tphcm
    -0.90
     pertenecen
    -0.90
    stücks
    -0.88
    POSITIVE LOGITS
     partially
    0.92
    ¢
    0.89
     with
    0.86
    ″
    0.85
     wholly
    0.85
    ornada
    0.83
     pObj
    0.83
     maar
    0.81
     simply
    0.81
    ]
    0.80
    Activations Density 0.022%

    No Known Activations