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

    Neuronpedia

    APIAssistant AxisNEWCircuit 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. 5659
    Prev
    Next
    INDEX
    Explanations

    technical terms and concepts related to data serialization and streams in programming

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

    serialization and byte array operations

    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
     InputDecoration
    -0.38
    новниш
    -0.36
     gobier
    -0.36
    berdayakan
    -0.36
     rêver
    -0.36
     stimmt
    -0.35
     Wird
    -0.34
    }{*}{
    -0.34
    lights
    -0.34
     lights
    -0.33
    POSITIVE LOGITS
    ByteArray
    1.52
     ByteArray
    1.36
    toByteArray
    0.90
     byteArray
    0.87
    byteArray
    0.74
    CharArray
    0.71
    FailureListener
    0.63
    aneously
    0.57
    IntArray
    0.56
    exca
    0.55
    Activations Density 0.001%

    No Known Activations