© 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. 22-GEMMASCOPE-RES-131K
    4. 54522
    Prev
    Next
    INDEX
    Explanations

    code definition and implementation

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

    The neuron fires on the main action verbs in top‐of‐file or top‐of‐class documentation comments (e.g. “configures,” “implements,” “defines,” “contains,” “holds,” “is,” “subclass,” etc.), i.e. the verbs that summarize what the code component does.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/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
     highest
    -1.19
     biggest
    -1.16
     нет
    -1.13
    teclado
    -1.12
     lots
    -1.11
     bigger
    -1.09
    也没
    -1.05
     weinig
    -1.05
     arrivée
    -1.04
     spacious
    -1.03
    POSITIVE LOGITS
     eğer
    1.23
     traduzione
    1.14
    ogues
    1.14
     GRATU
    1.13
    icted
    1.13
     kõik
    1.13
     ajustable
    1.12
     bayan
    1.11
     sinyal
    1.11
    keyframes
    1.10
    Activations Density 0.057%

    No Known Activations