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. 16-GEMMASCOPE-RES-65K
    6. 53201
    Prev
    Next
    INDEX
    Explanations

    syntactical elements and patterns typically found in code, such as prototype methods and object-oriented structure

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

    code keywords after punctuation

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 16-gemmascope-res-65k
    Configuration
    google/gemma-scope-2b-pt-res/layer_16/width_65k/average_l0_128
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    65,536
    Data Type
    float32
    Hook Name
    blocks.16.hook_resid_post
    Hook Layer
    16
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    -
    -0.61
    ↵
    -0.60
    ’
    -0.54
    '
    -0.54
    The
    -0.54
    ,
    -0.52
     The
    -0.52
    .
    -0.52
      
    -0.51
    ↵↵
    -0.49
    POSITIVE LOGITS
    windowFixed
    1.00
     Савезне
    0.95
    QMetaType
    0.93
     Мексичка
    0.93
    "])
    
    0.93
    Portale
    0.92
     Италијани
    0.91
    ArrowToggle
    0.91
    LookAnd
    0.91
    uxxxx
    0.90
    Activations Density 0.192%

    No Known Activations