© 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. 3915
    Prev
    Next
    INDEX
    Explanations

    like humans

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

    Explanation of neuron 4 behavior: the main thing this neuron does is spot numeric “position” tokens (floating‐point numbers) that mark a token’s place or break in the document.

    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
     ২
    0.59
     माह
    0.54
    কে
    0.53
     zwar
    0.51
    বা
    0.49
    which
    0.48
     cerveau
    0.48
     marchio
    0.46
    Oops
    0.46
    果然
    0.45
    POSITIVE LOGITS
    ถี่
    0.64
    Integration
    0.62
    𝓪
    0.62
     하는
    0.61
     profundamente
    0.61
    ]。
    0.60
    IPLE
    0.59
    <unused2037>
    0.59
     tastefully
    0.58
     soooo
    0.58
    Activations Density 0.037%

    No Known Activations