© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    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. 98092
    Prev
    Next
    INDEX
    Explanations

    specific entity names and references, particularly in technical or formal contexts

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

    HTT, MR, cult, inding, numerical values

    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
    TagMode
    -0.67
    tagHelperRunner
    -0.63
    addGroup
    -0.44
     ProtoMessage
    -0.41
    ValueStyle
    -0.40
    IsContent
    -0.40
    atoare
    -0.40
    levering
    -0.39
    言い
    -0.38
     <<<<<<<<<<<<<<
    -0.38
    POSITIVE LOGITS
     szóci
    0.47
    StoryboardSegue
    0.45
    Źródło
    0.41
     Мексичка
    0.40
     سكانية
    0.40
    HttpClientModule
    0.40
     GenerationType
    0.39
     normalize
    0.39
    glTranslatef
    0.38
     
    0.37
    Activations Density 0.158%

    No Known Activations