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 - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 13649
    Prev
    Next
    INDEX
    Explanations

    phrases emphasizing conditional or relative clauses

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    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
     itſelf
    -0.84
     Pelop
    -0.72
     alfo
    -0.71
     Houſe
    -0.71
     kasarigan
    -0.70
     Mahomet
    -0.69
     Shakspeare
    -0.69
     himſelf
    -0.67
     Anſ
    -0.67
     uchiha
    -0.67
    POSITIVE LOGITS
     there
    1.52
     they
    1.39
     the
    1.32
     it
    1.32
     we
    1.27
     this
    1.10
     you
    1.04
     he
    1.03
     a
    0.92
     some
    0.89
    Activations Density 0.365%

    No Known Activations