© 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. 2893
    Prev
    Next
    INDEX
    Explanations

    feature descriptions

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

    The neuron fires strongly on the first significant word of a new segment or heading (i.e. the initial content word marking the start of a section).

    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
    ,
    0.72
    ی
    0.69
    ைக
    0.65
    اظ
    0.65
    م
    0.65
    enso
    0.64
     '
    0.62
    ori
    0.62
     изда
    0.62
    —
    0.62
    POSITIVE LOGITS
     nameless
    0.86
    ssa
    0.78
     đô
    0.76
    OTROS
    0.75
    ピース
    0.75
     unequal
    0.75
    མ་
    0.72
    bbb
    0.71
     Parece
    0.71
     consequently
    0.71
    Activations Density 0.000%

    No Known Activations