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. 24349
    Prev
    Next
    INDEX
    Explanations

    references to programming or software artifacts

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

    \protect\mathpalette{\protect\independenT} notation

    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
     doccia
    -0.42
     total
    -0.42
     assoluto
    -0.40
    ูล
    -0.40
     fancy
    -0.37
     ArrayAdapter
    -0.35
     super
    -0.35
    的最大
    -0.35
    toMatch
    -0.34
     favor
    -0.34
    POSITIVE LOGITS
    Filmografie
    0.63
    richTextPanel
    0.57
    parsedMessage
    0.55
    ագրություններ
    0.54
    WebControls
    0.51
     queſta
    0.48
    లాలు
    0.47
    Viitteet
    0.47
    Datuak
    0.46
    Архівовано
    0.46
    Activations Density 0.000%

    No Known Activations