© 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. 24-GEMMASCOPE-2-RES-16K
    7. 3981
    Prev
    Next
    INDEX
    Explanations

    create a

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

    This neuron responds to instructional cue words that introduce questions or assignments—imperative prompt verbs (e.g. “Explain,” “Describe,” “Discuss,” “Give”) and question words (“Why,” “What,” “How,” etc.).

    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_24_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
     graines
    0.78
    이니까
    0.73
     因為
    0.72
     因为
    0.72
     입력
    0.70
    ０
    0.68
    𝕟
    0.67
     veux
    0.66
    ᅱ
    0.65
    nější
    0.65
    POSITIVE LOGITS
     the
    0.78
     exhibit
    0.67
    sthe
    0.67
     teh
    0.66
     three
    0.65
     briefly
    0.65
     THREE
    0.64
     structural
    0.63
     liaison
    0.63
     justification
    0.62
    Activations Density 0.005%

    No Known Activations