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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsBlogSlackPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-9B-IT
    4. Residual Stream - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 3426
    Prev
    Next
    INDEX
    Explanations

    punctuation and formatting elements, particularly dashes and hyphens

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91
    Prompts (Dashboard)
    24,576 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
    WriteLiteral
    -0.53
     للمعارف
    -0.46
     Helm
    -0.45
     Salary
    -0.45
    Malley
    -0.45
    ंदीखरीदारी
    -0.44
    Biografie
    -0.43
     autorytatywna
    -0.42
     Father
    -0.41
     Lemon
    -0.41
    POSITIVE LOGITS
     lisäksi
    0.54
     kysy
    0.54
     nemlig
    0.52
     elä
    0.51
    whether
    0.49
    namely
    0.49
     itulah
    0.48
     gebou
    0.48
    ​—
    0.48
     iaitu
    0.47
    Activations Density 0.291%

    No Known Activations