© 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
    2. Gemma-3-12B
    3. 31-GEMMASCOPE-2-RES-16K
    4. 13244
    Prev
    Next
    INDEX
    Explanations

    The neuron fires strongly on single-word descriptive modifiers (adjectives) that label types or categories (e.g. “raw” in “raw materials,” “spider” in “spider veins,” “narrow” in “narrow boat”).

    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
    5
    0.46
    वाहक
    0.46
    football
    0.46
    basketball
    0.44
    Span
    0.44
    4
    0.44
     League
    0.43
    ️
    0.43
    라이
    0.42
    Revolution
    0.42
    POSITIVE LOGITS
     인한
    0.55
     connectedness
    0.50
     deleterious
    0.48
     momentous
    0.48
     furthering
    0.47
    amination
    0.46
    ialmente
    0.46
     ambiguities
    0.46
     stateMutability
    0.46
     ascertaining
    0.45
    Activations Density 0.128%

    No Known Activations