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-2-2B
    4. Residual Stream - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 9345
    Prev
    Next
    INDEX
    Explanations

    concepts related to slowing down, calming, and narrowing focus

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

    deceleration or slowing

    np_acts-logits-general · gemini-2.0-flash

    slow down

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71
    Prompts (Dashboard)
    36,864 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
    XmlType
    -0.58
    clientWidth
    -0.58
     Eggers
    -0.54
    utico
    -0.53
    AddWithValue
    -0.53
     lí
    -0.52
    arcas
    -0.52
    出版年
    -0.52
    baomidou
    -0.51
    tvguidetime
    -0.51
    POSITIVE LOGITS
     down
    3.04
    down
    2.48
    Down
    2.38
     Down
    2.25
     DOWN
    2.14
    DOWN
    2.03
    downs
    1.93
     downs
    1.72
    ダウン
    1.67
    Downs
    1.48
    Activations Density 0.152%

    No Known Activations