© 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-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 92955
    Prev
    Next
    INDEX
    Explanations

    @ mentions

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

    This neuron detects tokens involved in “mentioning” or “addressing” someone or something—especially the @ symbol, the word “mention,” and related terms like channel or ID.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Selon
    -0.83
     Wich
    -0.82
     EINVAL
    -0.74
    imide
    -0.73
    apun
    -0.72
    jskiej
    -0.72
    ≡
    -0.71
    íř
    -0.71
     explan
    -0.70
     roboc
    -0.70
    POSITIVE LOGITS
     mention
    2.61
     @
    2.28
     mentions
    2.27
     tagging
    2.14
     tagged
    2.14
    mention
    2.14
     mentioned
    2.13
    Mention
    2.11
     Mention
    2.08
     mentioning
    1.99
    Activations Density 0.017%

    No Known Activations