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. MLP Out - 65k
    5. 0-GEMMASCOPE-MLP-65K
    6. 61077
    Prev
    Next
    INDEX
    Explanations

    repeated sequences of characters, specifically the string "jj"

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 0-gemmascope-mlp-65k
    Configuration
    google/gemma-scope-2b-pt-mlp/layer_0/width_65k/average_l0_72
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    65,536
    Data Type
    float32
    Hook Name
    blocks.0.hook_mlp_out
    Hook Layer
    0
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Monfieur
    -0.71
     againſt
    -0.70
     choice
    -0.68
     pleaſure
    -0.68
     purpoſe
    -0.67
     ſtate
    -0.66
     Eſ
    -0.65
     paſſ
    -0.65
     ſeveral
    -0.65
     ſtre
    -0.64
    POSITIVE LOGITS
     betweenstory
    1.19
    ViewFeatures
    0.86
    theless
    0.86
     Rptr
    0.73
    __':
    
    0.72
    etheless
    0.71
     NSCoder
    0.71
     endwhile
    0.69
    Literatuur
    0.69
     MainAxisSize
    0.68
    Activations Density 0.022%

    No Known Activations