© 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
    2. Gemma-2-27B
    3. 10-GEMMASCOPE-RES-131K
    4. 59046
    Prev
    Next
    INDEX
    Explanations

    proximity to landmarks

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

    The neuron specializes in flagging the functional “feature‐statement” language in a hotel or real‐estate listing—i.e. the verbs and function words (has, features, offers, within, from, a, the, and, etc.) that introduce location details, distances, and amenity lists.

    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
     but
    -1.66
     when
    -1.63
     they
    -1.58
     it
    -1.58
     will
    -1.50
    6
    -1.39
     because
    -1.37
    しましたが
    -1.23
     Unterneh
    -1.22
     regalías
    -1.20
    POSITIVE LOGITS
    کہ
    1.31
    NIP
    1.19
     of
    1.16
    ∽
    1.15
     fût
    1.15
     esas
    1.13
     comprends
    1.13
     Soit
    1.13
     onbe
    1.13
     dah
    1.10
    Activations Density 0.038%

    No Known Activations