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. 3698
    Prev
    Next
    INDEX
    Explanations

    structured data or elements related to reports and calculations

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

    Japanese text or code markers

    np_acts-logits-general · gemini-2.0-flash

    Japanese and English phrases

    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
     فريبيس
    -0.86
     <=",
    -0.70
    unnitel
    -0.67
    ImageContext
    -0.66
    arakat
    -0.64
     Normdatei
    -0.64
    UAWEI
    -0.63
    CrossRef
    -0.62
    новништво
    -0.60
    Geografi
    -0.60
    POSITIVE LOGITS
     же
    0.75
     дописавши
    0.71
    "]);
    
    0.69
    "];
    
    0.63
     doch
    0.58
    hogy
    0.58
    invokeLater
    0.56
    ']);
    
    0.55
    ]);
    
    0.54
    >");
    
    0.54
    Activations Density 0.105%

    No Known Activations