© 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. 5617
    Prev
    Next
    INDEX
    Explanations

    The neuron fires on subjective evaluative language—particularly adjectives and adverbs expressing opinions or assessments (e.g. unexpected, athletic, upright, more).

    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
    積分
    0.57
     PLATE
    0.56
     Trời
    0.56
     Verbal
    0.53
    ිනි
    0.52
    pons
    0.52
    والفق
    0.51
    🚽
    0.51
     dioxygen
    0.50
    PRODUCTS
    0.50
    POSITIVE LOGITS
     estilo
    1.51
     стиля
    1.48
     genre
    1.44
     genres
    1.42
    风格
    1.42
     styles
    1.41
     style
    1.39
     neoclassical
    1.38
     classical
    1.36
     gothic
    1.36
    Activations Density 0.993%

    No Known Activations