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-2-9B
    4. Residual Stream - 16k
    5. 21-GEMMASCOPE-RES-16K
    6. 11556
    Prev
    Next
    INDEX
    Explanations

    structured data types and programming-related elements

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B @ 21-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-pt-res/layer_21/width_16k/average_l0_129
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.21.hook_resid_post
    Hook Layer
    21
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     ailleurs
    -0.41
    昏
    -0.40
    BRON
    -0.40
    Décès
    -0.39
     stof
    -0.38
     NSError
    -0.38
    lihatan
    -0.38
    \{\\
    -0.38
     ensued
    -0.38
    IRON
    -0.37
    POSITIVE LOGITS
    ">//
    0.59
     informée
    0.52
    complexContent
    0.44
    verifyException
    0.44
     تضيفلها
    0.41
     endblock
    0.40
    ScopeManager
    0.38
     мәкал
    0.38
    UpdateCommand
    0.37
     Unmarshal
    0.37
    Activations Density 0.024%

    No Known Activations