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. 11099
    Prev
    Next
    INDEX
    Explanations

    questions and interactions related to discussions or dialogues

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

    Following numbers in questions

    np_acts-logits-general · gemini-2.0-flash

    numbers and math questions

    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-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
    ^(@)
    -1.11
    ++
    
    -1.00
    ".
    
    -0.99
    ]--;
    -0.91
    %");
    -0.91
     itſelf
    -0.90
    */;
    -0.89
    }.
    
    -0.86
     $_"
    -0.86
     ་་
    -0.86
    POSITIVE LOGITS
     I
    1.20
    I
    0.97
     nice
    0.86
     ok
    0.81
     my
    0.81
     What
    0.78
     You
    0.78
     okay
    0.78
     Maybe
    0.78
    ok
    0.77
    Activations Density 1.271%

    No Known Activations