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. 120174
    Prev
    Next
    INDEX
    Explanations

    instances of punctuation and sentence-ending phrases

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

    fast, but, however

    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
     autorytatywna
    -0.78
     nahilalakip
    -0.64
    Geografi
    -0.57
    hyrchwyd
    -0.56
    setArguments
    -0.55
    RegressionTest
    -0.53
     ujednoznacz
    -0.53
    fromnode
    -0.53
     AssemblyTitle
    -0.51
    Чыгана
    -0.51
    POSITIVE LOGITS
    imageshack
    0.45
     added
    0.36
     הב
    0.36
     nonatomic
    0.34
    ecap
    0.33
    OrNil
    0.32
    CppMethod
    0.31
     actionPerformed
    0.31
     adds
    0.31
    siede
    0.31
    Activations Density 0.020%

    No Known Activations