© 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. 447
    Prev
    Next
    INDEX
    Explanations

    starting sections with descriptive titles

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

    This neuron activates on words in the body of a hard-news wire report (e.g. the dateline and formal reporting text).

    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
     you
    -1.43
     I
    -1.43
     my
    -1.39
    ׇ
    -1.27
    こんな感じ
    -1.23
     Erzb
    -1.20
     litri
    -1.18
     wonderful
    -1.17
    那些
    -1.13
    只要
    -1.11
    POSITIVE LOGITS
    ּוֹ
    2.08
    ַּ
    1.71
    וּ
    1.70
     “[
    1.55
    ְּ
    1.48
     Unii
    1.41
     грамм
    1.38
    🪛
    1.36
     Verwenden
    1.34
    ֔
    1.32
    Activations Density 0.020%

    No Known Activations