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 - 16k
    5. 18-GEMMASCOPE-RES-16K
    6. 8770
    Prev
    Next
    INDEX
    Explanations

    keywords and concepts related to death, loss, and crime

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

    Chinese characters before certain nouns

    np_acts-logits-general · gemini-2.0-flash

    death procedure

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 18-gemmascope-res-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_18/width_16k/average_l0_74
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.18.hook_resid_post
    Hook Layer
    18
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ,
    -0.81
     
    -0.72
    .
    -0.68
     of
    -0.66
      
    -0.62
     in
    -0.61
     (
    -0.61
     -
    -0.60
     I
    -0.59
     \
    -0.59
    POSITIVE LOGITS
    )");
    
    1.18
     Reſ
    1.13
     $_"
    1.13
    )";
    
    1.11
     auffi
    1.11
     ExecuteAsync
    1.07
     ſeveral
    1.07
    expandindo
    1.06
    )"),
    1.06
     itſelf
    1.04
    Activations Density 0.067%

    No Known Activations