Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsBlogSlackPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-9B
    4. Residual Stream - 131k
    5. 20-GEMMASCOPE-RES-131K
    6. 53876
    Prev
    Next
    INDEX
    Explanations

    actions and changes occurring over time, often depicted as gradual or slow processes

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-9b-pt-res/layer_20/width_131k/average_l0_114
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    131,072
    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
     ricc
    -0.42
     currency
    -0.41
     polarity
    -0.40
     wildest
    -0.39
     challenge
    -0.39
     rique
    -0.38
    HIGH
    -0.38
     berke
    -0.37
     HIGH
    -0.36
     Challenge
    -0.36
    POSITIVE LOGITS
    Slowly
    1.12
     slowly
    1.08
     Slowly
    1.06
    slowly
    1.05
     gradually
    0.92
     Gradually
    0.92
     perlahan
    0.89
     langsam
    0.82
     lentamente
    0.82
     lentement
    0.82
    Activations Density 0.023%

    No Known Activations