© 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. 120017
    Prev
    Next
    INDEX
    Explanations

    in many ways

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

    This neuron specifically fires on the word “ways” when it appears in idiomatic phrases (e.g. “in many ways,” “in some ways,” “in a lot of ways,” etc.).

    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
    🫔
    -1.75
    ྜ
    -1.75
    Ꮵ
    -1.72
    ?
    
    -1.64
    杳
    -1.59
    )+"
    -1.59
    Това
    -1.58
    𓋼
    -1.58
    ственность
    -1.55
    頔
    -1.54
    POSITIVE LOGITS
    «
    1.90
     \
    1.80
     we
    1.67
     sahen
    1.64
    en
    1.62
    AutoresizingMask
    1.60
    béco
    1.59
     других
    1.57
     ένα
    1.52
    ti
    1.50
    Activations Density 0.011%

    No Known Activations