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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & 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. 16-GEMMASCOPE-RES-65K
    6. 53412
    Prev
    Next
    INDEX
    Explanations

    symbols and mathematical constructs used in formal expressions

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

    mathematical concepts and structures

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 16-gemmascope-res-65k
    Configuration
    google/gemma-scope-2b-pt-res/layer_16/width_65k/average_l0_128
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    65,536
    Data Type
    float32
    Hook Name
    blocks.16.hook_resid_post
    Hook Layer
    16
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ọa
    -0.54
     West
    -0.51
    İstinadlar
    -0.47
     Am
    -0.47
    pady
    -0.46
     Sec
    -0.45
     padx
    -0.45
    fflush
    -0.45
    óricas
    -0.45
    lojik
    -0.45
    POSITIVE LOGITS
    ValueStyle
    0.75
     itſelf
    0.75
    ########.
    0.65
    Portail
    0.65
     Chwiliwch
    0.62
    SourceChecksum
    0.61
     autorytatywna
    0.61
     concept
    0.60
     organizations
    0.60
     MERCHANTABILITY
    0.59
    Activations Density 0.208%

    No Known Activations