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-IT
    4. Residual Stream - 131k
    5. 20-GEMMASCOPE-RES-131K
    6. 30219
    Prev
    Next
    INDEX
    Explanations

    references to basketball and related activities

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

    basketball, NBA, players, and tournaments

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k
    Configuration
    google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81
    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
    躇
    -0.57
    setopt
    -0.48
    endforeach
    -0.45
     Deixe
    -0.44
     potvr
    -0.43
    °)
    -0.42
    foy
    -0.42
     dök
    -0.42
     cuarzo
    -0.41
     iniziale
    -0.41
    POSITIVE LOGITS
     basketball
    1.52
     Basketball
    1.48
    basketball
    1.42
    Basketball
    1.41
     BASKET
    0.96
    🏀
    0.92
    篮球
    0.89
     NBA
    0.87
     baloncesto
    0.85
     Basket
    0.80
    Activations Density 0.007%

    No Known Activations