© 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. 31-GEMMASCOPE-2-RES-16K
    7. 2638
    Prev
    Next
    INDEX
    Explanations

    caus, reso, templ, opin, writ, ref, gamb, look

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

    This neuron activates strongly on uncommon or low‐frequency subword tokens (such as rare prefixes, suffixes, or proper‐name fragments).

    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_31_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
    ের
    0.82
    schnitt
    0.79
    েরিক
    0.69
     乾燥
    0.68
    য়
    0.68
    の
    0.67
    ገኛ
    0.66
    LDL
    0.66
     Cancer
    0.65
    ម្លៃ
    0.64
    POSITIVE LOGITS
    াসহ
    0.81
    sthrough
    0.74
    ­
    0.68
    sthe
    0.67
    ̓
    0.67
    ें
    0.66
    水印
    0.65
    ੍
    0.65
    utenants
    0.64
    ableness
    0.63
    Activations Density 1.446%

    No Known Activations