© 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
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 39136
    Prev
    Next
    INDEX
    Explanations

    specific concepts and outcomes

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

    The neuron activates on single-word abstract concepts or measured quantities (e.g. “evidence,” “details,” “numbers,” “consequences”), picking out nouns that express ideas or metrics rather than concrete objects.

    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
    selben
    -2.00
    (
    -1.88
    :
    -1.77
    7
    -1.75
    로
    -1.67
    蕺
    -1.65
    8
    -1.54
    𝒕
    -1.54
     bemerk
    -1.53
    5
    -1.53
    POSITIVE LOGITS
     there
    2.25
     has
    1.78
     through
    1.70
     they
    1.69
     Without
    1.63
    犼
    1.63
     says
    1.62
     tells
    1.62
     purposefully
    1.60
     wants
    1.58
    Activations Density 0.121%

    No Known Activations