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 - 65k
    5. 20-GEMMASCOPE-RES-65K
    6. 600
    Prev
    Next
    INDEX
    Explanations

    numerical values or references

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

    indicators of significant statistical events or findings in a document

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

    numbers and years

    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-65k
    Configuration
    google/gemma-scope-2b-pt-res/layer_20/width_65k/average_l0_114
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    65,536
    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
     demain
    -0.73
     today
    -0.70
     yesterday
    -0.66
     Yesterday
    -0.66
     amanhã
    -0.65
     idag
    -0.65
    yesterday
    -0.65
     tonight
    -0.65
    today
    -0.65
     σήμερα
    -0.64
    POSITIVE LOGITS
    sizeCache
    0.65
     ComVisible
    0.61
     RoutedEventArgs
    0.60
    Geplaatst
    0.58
     when
    0.57
    veratrol
    0.57
    SOUNDBITE
    0.57
    campal
    0.56
    هنگ
    0.53
    IntoConstraints
    0.51
    Activations Density 0.116%

    No Known Activations