© 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. 79700
    Prev
    Next
    INDEX
    Explanations

    specific nountopics

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

    The neuron activates on capitalized words that are part of formal names or titles (e.g. products, campaigns, books, pages, or services).

    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
    5
    -1.50
    6
    -1.45
    9
    -1.42
    8
    -1.42
    7
    -1.24
    2
    -1.16
    4
    -1.15
     other
    -1.11
     time
    -1.11
     料理
    -1.05
    POSITIVE LOGITS
     吗
    1.58
    )))),
    1.48
    念头
    1.43
    PointerException
    1.38
    Frequently
    1.38
    
    1.38
    珧
    1.38
     kteří
    1.37
    yelitis
    1.37
    ihnachts
    1.35
    Activations Density 0.191%

    No Known Activations