© 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. 3186
    Prev
    Next
    INDEX
    Explanations

    following sentence or text

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

    This neuron detects the formatting and keywords of numbered question prompts and instruction headings (e.g. “Directions,” question numbers, and related labels).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    instructional text that directs readers to complete tasks, answer questions, or follow specific procedures.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @jyhe0408

    instructions and directions for exercises or questions in educational/worksheet contexts, especially numbered sections and prompts to review, select, match, or answer items.

    oai_token-act-pair · gpt-5Triggered 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
    ___________
    0.72
     ________
    0.69
    ____________
    0.67
     ____________
    0.66
     ______
    0.66
     _______
    0.66
    ________
    0.65
     collim
    0.65
     __________
    0.64
     electrons
    0.63
    POSITIVE LOGITS
     seguenti
    1.02
    以下の
    1.02
     seguente
    1.01
    以下
    1.00
    following
    0.96
     følgende
    0.95
     siguientes
    0.95
     following
    0.92
    下記
    0.92
     folgenden
    0.91
    Activations Density 0.017%

    No Known Activations