Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & 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. 22-GEMMASCOPE-RES-16K
    6. 10313
    Prev
    Next
    INDEX
    Explanations

    discussions of specific scenarios and critical points in a variety of contexts

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

    Followed by "where"

    np_acts-logits-general · gemini-2.0-flash

    scenarios where contexts occur

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 22-gemmascope-res-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_22/width_16k/average_l0_72
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.22.hook_resid_post
    Hook Layer
    22
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     nahilalakip
    -0.65
    windowFixed
    -0.61
    arschijnlijk
    -0.58
    ---|---|
    -0.57
    BorderFactory
    -0.54
     ProtoMessage
    -0.52
     &___
    -0.50
    kelijkheid
    -0.49
    それと
    -0.48
    theless
    -0.46
    POSITIVE LOGITS
     where
    4.36
    where
    3.75
    Where
    3.14
     Where
    3.07
     donde
    3.01
     où
    2.94
     WHERE
    2.92
     где
    2.63
    WHERE
    2.60
    donde
    2.43
    Activations Density 2.179%

    No Known Activations