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 - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 10289
    Prev
    Next
    INDEX
    Explanations

    technical descriptions of components in circuits

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

    Components or parts of a device/circuit

    np_acts-logits-general · gemini-2.0-flash

    technical components and circuits

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.20.hook_resid_post
    Hook Layer
    20
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
      
    -0.51
     oficial
    -0.50
    <em>
    -0.50
     dodgy
    -0.50
     unofficial
    -0.47
     Undeclared
    -0.46
     seemed
    -0.46
     seems
    -0.45
     chimneys
    -0.45
     shady
    -0.45
    POSITIVE LOGITS
     thereto
    1.02
     plurality
    1.00
     therefrom
    1.00
     substantially
    0.99
     thereof
    0.96
    wherein
    0.95
     wherein
    0.92
     therein
    0.90
     Preferably
    0.87
     aforesaid
    0.85
    Activations Density 2.686%

    No Known Activations