© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 11080
    Prev
    Next
    INDEX
    Explanations

    prepend with backslash sequence

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

    The neuron detects garbled or mis-decoded text sequences (mojibake), such as stray “â,” “€,” or similar artifact characters.

    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
    <unused98>
    0.64
     cyclohex
    0.62
     상세보기
    0.61
     séjour
    0.59
     якості
    0.59
    zeugt
    0.58
     způso
    0.58
     varphi
    0.58
    হেতু
    0.57
     phenyl
    0.57
    POSITIVE LOGITS
    „
    0.81
     „
    0.79
    ”
    0.71
    ，「
    0.66
    「
    0.66
    “
    0.66
    "
    0.64
    Community
    0.63
     ‚
    0.62
    .“
    0.61
    Activations Density 0.047%

    No Known Activations