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

    Neuronpedia

    APICircuit 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. 20-GEMMASCOPE-RES-65K
    6. 958
    Prev
    Next
    INDEX
    Explanations

    verbal phrases indicating actions or emotions

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

    expressions of contradiction or hypocrisy in political contexts

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

    to the extreme

    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
     autorytatywna
    -0.49
    enderror
    -0.48
    withIdentifier
    -0.47
    HtmlAttribute
    -0.47
    CppCodeGen
    -0.45
    phazard
    -0.45
    Aholisi
    -0.44
    گران
    -0.44
    Unmarshaller
    -0.44
     AssemblyProduct
    -0.44
    POSITIVE LOGITS
     hasta
    0.97
     till
    0.91
    beyond
    0.88
     até
    0.87
     sampai
    0.87
     extreme
    0.85
     beyond
    0.85
     jusqu
    0.81
    極
    0.81
     max
    0.81
    Activations Density 0.140%

    No Known Activations