© 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. 5902
    Prev
    Next
    INDEX
    Explanations

    can test | Nation clause | final IF | Emergency Reload

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

    This neuron strongly activates on individual segments of hyphenated compound words (e.g. the parts appearing immediately before or after a “-”).

    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
    💩
    0.60
    Classpath
    0.58
    〰
    0.58
    RECE
    0.57
    Gamepad
    0.57
    RFID
    0.57
    ⏱
    0.56
    🤜
    0.56
    🎸
    0.56
    PUBLIC
    0.55
    POSITIVE LOGITS
    と呼ばれる
    0.84
     waiver
    0.72
     caso
    0.68
     globals
    0.66
     Special
    0.62
     clause
    0.61
     Category
    0.61
     globo
    0.61
     syndrome
    0.60
     Memorandum
    0.60
    Activations Density 0.469%

    No Known Activations