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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & 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. 42367
    Prev
    Next
    INDEX
    Explanations

    technical terms and jargon related to specific fields or specialized topics

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

    Silver, kohlrabi, cholangiography

    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
    fastjson
    -0.47
     bpy
    -0.42
    ftagPool
    -0.38
     Legendre
    -0.38
     agriculture
    -0.38
     assemblies
    -0.37
     aggregation
    -0.37
    näm
    -0.37
    WithData
    -0.37
    celotti
    -0.37
    POSITIVE LOGITS
     eyeliner
    0.71
    eliner
    0.63
     الرياضيه
    0.52
    principalColumn
    0.52
    +#+#
    0.51
     للاسماء
    0.46
    roek
    0.45
     surla
    0.44
     autorytatywna
    0.44
     @"/
    0.44
    Activations Density 0.001%

    No Known Activations