© 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
    4. Residual Stream - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 9621
    Prev
    Next
    INDEX
    Explanations

    technical terms and diagnostic procedures related to troubleshooting

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

    text describing technical testing, troubleshooting, or diagnostic procedures.

    oai_token-act-pair · claude-3-7-sonnet-20250219Triggered by @neilrathi

    language about troubleshooting and diagnostic testing—checking conditions, using tools or measurements, observing responses, and asking evaluative questions.

    oai_token-act-pair · gpt-5Triggered by @jyhe0408

    words related to diagnostic testing and troubleshooting procedures.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B @ 20-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-pt-res/layer_20/width_16k/average_l0_68
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    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
    openzeppelin
    -0.51
    énario
    -0.47
    rawValue
    -0.46
     desmotivaciones
    -0.46
    orologio
    -0.41
     laço
    -0.41
    awatan
    -0.40
     OFDb
    -0.40
    LAYER
    -0.39
    綜合報導
    -0.39
    POSITIVE LOGITS
     otomatig
    0.47
     للمعارف
    0.47
    WithMany
    0.47
     snippetHide
    0.42
    delwed
    0.41
    MigrationBuilder
    0.41
     <>",
    0.41
    multirow
    0.41
    تقاوى
    0.38
     verify
    0.37
    Activations Density 0.155%

    No Known Activations