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. 29605
    Prev
    Next
    INDEX
    Explanations

    expressions of surprise or shock in the text

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

    surprise, surprising, surprisingly

    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
    Життєпис
    -0.43
     Lauderdale
    -0.40
     TDA
    -0.40
    要
    -0.40
     Bootcamp
    -0.39
     angelegt
    -0.39
    Ged
    -0.39
    荫
    -0.38
     queremos
    -0.38
    frontmatter
    -0.38
    POSITIVE LOGITS
     surprising
    0.77
     éton
    0.74
     surprised
    0.73
     sorprend
    0.73
     erstaun
    0.71
     astonished
    0.67
    surprised
    0.67
     Surprisingly
    0.66
     surpren
    0.66
     surprisingly
    0.65
    Activations Density 0.016%

    No Known Activations