Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-2B
    4. Residual Stream - 65k
    5. 20-GEMMASCOPE-RES-65K
    6. 126
    Prev
    Next
    INDEX
    Explanations

    marker phrases indicating the beginning of a section or block of content

    oai_token-act-pair · gpt-4o-miniTriggered by @bot

    the start of various sections or segments in a script or code

    oai_token-act-pair · gpt-4o-miniTriggered by @bot

    special tool set

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 20-gemmascope-res-65k
    Configuration
    google/gemma-scope-2b-pt-res/layer_20/width_65k/average_l0_114
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    65,536
    Data Type
    float32
    Hook Name
    blocks.20.hook_resid_post
    Hook Layer
    20
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    richTextPanel
    -0.80
     <<<<<<<<<<<<<<
    -0.78
    ArgsConstructor
    -0.77
    TagMode
    -0.74
     resourceCulture
    -0.71
     AssemblyCulture
    -0.68
    ineno
    -0.67
    Personendaten
    -0.66
    webElementXpaths
    -0.66
    ///</
    -0.66
    POSITIVE LOGITS
     همچنین
    0.55
     crdi
    0.53
    همچنین
    0.48
    otriva
    0.47
    特有
    0.47
     также
    0.47
     مواليد
    0.47
    nande
    0.47
     asimismo
    0.46
     alfo
    0.44
    Activations Density 0.019%

    No Known Activations