© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-2-27B
    3. 10-GEMMASCOPE-RES-131K
    4. 131018
    Prev
    Next
    INDEX
    Explanations

    research study descriptions

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

    The neuron activates strongly for nouns related to documents, publications, or formal written content such as "book," "dish," "house," "company," "table," "study," "page," and "matter."

    oai_token-act-pair · claude-3-7-sonnet-20250219Triggered by @neilrathi

    The neuron strongly activates on the main head nouns in formal or descriptive phrases (e.g. “COMPANY,” “house,” “soul,” “outcome,” “study”)—in other words, it spots the principal noun in section headings or key thematic statements.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     have
    -2.13
     from
    -2.02
     in
    -1.99
     one
    -1.89
     a
    -1.89
     also
    -1.84
     of
    -1.82
     an
    -1.76
     with
    -1.73
     had
    -1.61
    POSITIVE LOGITS
    翃
    1.73
     bascul
    1.65
     absolut
    1.63
     グラフ
    1.63
     écrire
    1.63
     fournisseur
    1.58
    ■
    1.58
    焿
    1.56
    牙齿
    1.55
    ◆◆
    1.55
    Activations Density 0.160%

    No Known Activations