© 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. 1867
    Prev
    Next
    INDEX
    Explanations

    message fragments, communications

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

    This neuron specifically detects occurrences of the word “message” (and its close variants).

    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
    shit
    0.75
    गर्
    0.60
    طاء
    0.59
    ն
    0.58
     Johnston
    0.56
    CERT
    0.56
    ECUT
    0.55
    jec
    0.55
    貨
    0.54
    Cub
    0.54
    POSITIVE LOGITS
     Message
    1.44
     msg
    1.36
    メッセージ
    1.24
     message
    1.23
     mensaje
    1.23
     메시
    1.23
     Messages
    1.23
     mesaj
    1.19
     messages
    1.17
     messaggio
    1.17
    Activations Density 0.231%

    No Known Activations