© 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. 12736
    Prev
    Next
    INDEX
    Explanations

    names and positive superlatives

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

    The neuron activates on numeric expressions with multipliers and units (e.g. race‐distance notations like “4 × 5 kilometre”).

    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
     in
    -1.56
     new
    -1.48
     after
    -1.40
     one
    -1.34
     before
    -1.32
     first
    -1.29
     when
    -1.28
     While
    -1.28
     while
    -1.27
     time
    -1.26
    POSITIVE LOGITS
     astounding
    1.45
     AMAZING
    1.36
    嘮
    1.23
     amazingly
    1.22
     verhind
    1.20
     unbelievably
    1.20
     probablement
    1.19
     magnificent
    1.19
     réparer
    1.18
     poë
    1.18
    Activations Density 0.013%

    No Known Activations