© 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. 8431
    Prev
    Next
    INDEX
    Explanations

    image files, developing, degradation

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

    The neuron fires strongly on relational prepositions and connective verbs (e.g. “with,” “to,” “contact,” “served,” “revelation”) that introduce objects or purposes.

    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
     Eugene
    0.86
    тна
    0.83
     McGu
    0.82
    ྥ
    0.82
    _<
    0.79
    珹
    0.79
    تنا
    0.79
     Sér
    0.78
     Sood
    0.78
     Geoff
    0.78
    POSITIVE LOGITS
    c
    0.98
    cac
    0.87
    f
    0.84
    v
    0.78
    blue
    0.76
    ur
    0.74
    w
    0.73
    ve
    0.72
    crc
    0.72
    cypress
    0.72
    Activations Density 0.000%

    No Known Activations