© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 11412
    Prev
    Next
    INDEX
    Explanations

    file paths and code syntax

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

    The neuron strongly activates on substrings of web addresses (e.g. domain names, path/query tokens like “policies.google.com”, “hl=en”, “tools.google.com/dlpage”, “xlibris.is”), i.e. URL or hyperlink 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
    neſs
    0.52
     Ciò
    0.51
     {:
    0.48
    俸
    0.47
     Roshan
    0.45
    校长
    0.45
     EEA
    0.44
     হান
    0.44
    GameState
    0.44
    रेक्ट
    0.44
    POSITIVE LOGITS
    즉
    0.56
    Ŕ
    0.56
     brillante
    0.55
     wym
    0.54
     produite
    0.54
    اح
    0.53
     즉
    0.52
     varm
    0.52
     amplio
    0.51
     అయితే
    0.50
    Activations Density 0.296%

    No Known Activations