© 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-3-12B
    5. 24-GEMMASCOPE-2-RES-16K
    7. 1168
    Prev
    Next
    INDEX
    Explanations

    characters following letter codes

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

    The neuron activates primarily on isolated single-letter alphabetic tokens (e.g. standalone “L”, “O”, “V”, “E”) rather than on full words.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-pt/resid_post/layer_24_width_16k_l0_medium
    Prompts (Dashboard)
    392,802 prompts, 256 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    óle
    0.70
    v
    0.64
    ری
    0.61
    aj
    0.59
    jmath
    0.58
    ఇ
    0.58
    BleStatus
    0.57
    igraf
    0.57
    っいて
    0.56
    manager
    0.55
    POSITIVE LOGITS
     Britain
    0.56
     Amenities
    0.54
     ___
    0.50
     (
    0.50
     Mason
    0.49
     N
    0.49
     Consecutive
    0.49
     ∆
    0.48
     hugging
    0.48
     Sequential
    0.48
    Activations Density 0.010%

    No Known Activations