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. 8430
    Prev
    Next
    INDEX
    Explanations

    instances of specific stages, groups, and structures in data or processes

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

    second item/group

    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
     Савезне
    -0.53
    setVerticalGroup
    -0.49
     Paglinawan
    -0.46
     ComVisible
    -0.43
     EconPapers
    -0.43
     мәкал
    -0.43
    YourGuide
    -0.42
     Exactos
    -0.42
     становника
    -0.42
    AutoField
    -0.41
    POSITIVE LOGITS
    -------
    0.45
    HtmlAttribute
    0.44
    今度は
    0.41
    EDEFAULT
    0.41
    PropertyGroup
    0.41
    Second
    0.41
    ModelAdmin
    0.40
    reditation
    0.40
    UpInside
    0.39
     arrangements
    0.38
    Activations Density 0.076%

    No Known Activations