© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-3-12B
    3. 31-GEMMASCOPE-2-RES-16K
    4. 11313
    Prev
    Next
    INDEX
    Explanations

    This is

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

    The neuron fires strongly on discourse‐level demonstrative or referential pronouns (e.g. “this,” “that,” “these”) used to link or point back to earlier content.

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

    the demonstrative pronoun "that" (and sometimes "this") when used to refer back to previously mentioned ideas or situations.

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

    sentence-initial demonstrative references that point to or emphasize a statement or idea, often in “X is/was” constructions.

    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
     proprietor
    0.64
     enables
    0.63
     corresponds
    0.63
     permettent
    0.63
     variant
    0.61
     moiety
    0.61
     facilitates
    0.60
     संस्करण
    0.57
     tranche
    0.57
     allows
    0.57
    POSITIVE LOGITS
     dilemmas
    0.74
     philosophies
    0.73
     такими
    0.72
    思想
    0.69
     이런
    0.68
     подобных
    0.66
    这种情况
    0.64
    arguments
    0.63
    argument
    0.63
     مسائل
    0.63
    Activations Density 0.170%

    No Known Activations