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
    4. Residual Stream - 131k
    5. 20-GEMMASCOPE-RES-131K
    6. 103020
    Prev
    Next
    INDEX
    Explanations

    elements and attributes in XML or XSLT code

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

    Code syntax elements in programming languages, particularly property access notation, dictionary keys, array indices, and XML tag closures in programming contexts.

    eleuther_acts_top20 · claude-3-7-sonnet-20250219Triggered by @lu-christina
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-9b-pt-res/layer_20/width_131k/average_l0_114
    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
    AddTagHelper
    -0.92
    MessageTagHelper
    -0.88
     GenerationType
    -0.85
    IsMutable
    -0.85
    ViewImports
    -0.82
    parsedMessage
    -0.78
    expandindo
    -0.77
    ValueGeneration
    -0.77
     surla
    -0.76
     queſta
    -0.76
    POSITIVE LOGITS
     presente
    0.36
     Ganze
    0.34
     style
    0.31
     value
    0.31
     represented
    0.30
     context
    0.30
     specific
    0.29
    ∬
    0.28
    Context
    0.28
     particular
    0.28
    Activations Density 0.154%

    No Known Activations