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. 29897
    Prev
    Next
    INDEX
    Explanations

    references to size and scaling in a scientific context

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

    reducing, decreasing, narrowing

    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
    migrationBuilder
    -0.58
    MLLoader
    -0.44
    很大
    -0.42
    🖨
    -0.40
    LIBRA
    -0.40
     arbitrator
    -0.39
    Pir
    -0.39
     lepas
    -0.38
    mbalikan
    -0.38
    Aiheesta
    -0.37
    POSITIVE LOGITS
     tighter
    0.84
     tight
    0.73
    Narrow
    0.71
     narrow
    0.71
     tightened
    0.71
     narrower
    0.70
     Narrow
    0.69
     compact
    0.67
    tight
    0.65
     tighten
    0.65
    Activations Density 0.141%

    No Known Activations