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. 14494
    Prev
    Next
    INDEX
    Explanations

    elements related to unique abilities and characteristics

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

    followed by "make" or "makes"

    np_acts-logits-general · gemini-2.0-flash

    makes or allows

    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
    OrBuilder
    -0.62
     şə
    -0.54
    ftagPool
    -0.52
    {{$
    -0.51
     Políticas
    -0.50
    autoconfigure
    -0.50
    Begriffsklä
    -0.48
    Inno
    -0.47
    τουργ
    -0.47
    setItem
    -0.47
    POSITIVE LOGITS
     membuatnya
    1.07
    kannya
    0.99
    使其
    0.86
     makes
    0.77
     menjadikan
    0.70
    makes
    0.70
     puts
    0.70
     enabled
    0.70
    Makes
    0.69
     him
    0.69
    Activations Density 0.365%

    No Known Activations