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. 4990
    Prev
    Next
    INDEX
    Explanations

    elements of C++ programming code related to type definitions and protocols

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

    C++ related code

    np_acts-logits-general · gemini-2.0-flash

    static cast or reinterpret cast

    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
    \{\\
    -0.52
     SwitchCompat
    -0.52
     <<<<<<<<<<<<<<
    -0.48
    ofür
    -0.47
    LookAnd
    -0.44
    jenost
    -0.41
    ̈́
    -0.38
    marvin
    -0.37
     Hu
    -0.37
     يتيمه
    -0.37
    POSITIVE LOGITS
     dynamic
    1.12
    static
    1.10
     static
    1.09
    dynamic
    1.08
    Dynamic
    0.81
     dynamique
    0.81
     DYNAMIC
    0.81
     Dynamic
    0.80
    reinterpret
    0.78
     dinámico
    0.76
    Activations Density 0.201%

    No Known Activations