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. 130993
    Prev
    Next
    INDEX
    Explanations

    instances of the term "sold."

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

    A beginning of sequence token that marks the start of a new text document or passage in a corpus, appearing at natural document boundaries where the topic changes.

    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
    :✨
    -0.96
     ProtoMessage
    -0.89
     estekak
    -0.85
    Diweddarwch
    -0.82
    rungsseite
    -0.81
     snippetHide
    -0.80
    IntOverflow
    -0.78
    WireFormatLite
    -0.77
    RTEE
    -0.75
    ьажоргаш
    -0.75
    POSITIVE LOGITS
    Современ
    0.37
    ↵
    0.35
    Modern
    0.34
     
    0.34
    com
    0.33
    <strong>
    0.32
     Modern
    0.32
    _
    0.32
     pre
    0.32
    тори
    0.30
    Activations Density 0.076%

    No Known Activations