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 - 65k
    5. 17-GEMMASCOPE-RES-65K
    6. 28471
    Prev
    Next
    INDEX
    Explanations

    license and copyright text from open-source software headers.

    oai_token-act-pair · claude-3-5-haiku-20241022Triggered by @chuloivawork

    keywords and phrases associated with licensing and copyright

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 17-gemmascope-res-65k
    Configuration
    google/gemma-scope-2b-pt-res/layer_17/width_65k/average_l0_125
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    65,536
    Data Type
    float32
    Hook Name
    blocks.17.hook_resid_post
    Hook Layer
    17
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ^(@)
    -0.33
    tershire
    -0.31
    drawal
    -0.30
    encils
    -0.30
    ficulty
    -0.29
    %\]
    -0.29
    ]--;
    -0.29
    ^(@
    -0.28
    )++;
    -0.28
    pecially
    -0.27
    POSITIVE LOGITS
    <bos>
    0.96
    enumi
    0.40
     estadounid
    0.39
    存于互联网档案馆
    0.38
    </tr>
    0.36
    жде
    0.35
    MathML
    0.35
    ρυ
    0.33
     indígen
    0.33
     ſei
    0.32
    Activations Density 39.965%

    No Known Activations