© 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. 100521
    Prev
    Next
    INDEX
    Explanations

    oratory, orator, Horace, Bacon, cicero

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

    The neuron fires strongly on the initial tokens of section headers or document titles—e.g. names like “Seneca,” “Gorgias,” the Latin species epithet in an article title, or the `(ns …)`/`(:require …)` form in Clojure code—marking the start of a new document or major subsection.

    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
    gencias
    -1.81
    with
    -1.68
    そのような
    -1.62
    就像是
    -1.59
     ditt
    -1.58
    ꜆
    -1.55
    mbert
    -1.54
    d
    -1.53
     produisent
    -1.53
    hilterra
    -1.53
    POSITIVE LOGITS
     复古
    1.80
     tne
    1.72
     Cicero
    1.59
     入り口
    1.55
     histoires
    1.48
     écoles
    1.45
     từ
    1.45
    Meik
    1.44
     \&
    1.44
    .
    1.41
    Activations Density 0.012%

    No Known Activations