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. 20-GEMMASCOPE-RES-65K
    6. 1017
    Prev
    Next
    INDEX
    Explanations

    instances of the end-of-line command in code

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

    instances of the word "endl" typically used in C++ programming

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

    end of line followed by semicolon

    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-65k
    Configuration
    google/gemma-scope-2b-pt-res/layer_20/width_65k/average_l0_114
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    65,536
    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
    +#+#
    -0.70
    __(/*!
    -0.70
    SourceChecksum
    -0.62
    ا
    -0.58
    MigrationBuilder
    -0.56
    a
    -0.55
    o
    -0.51
    ض
    -0.51
    uza
    -0.49
    late
    -0.48
    POSITIVE LOGITS
     purpoſe
    0.88
     Cister
    0.82
     GIPHY
    0.77
     endl
    0.77
     iſt
    0.76
     Shakspeare
    0.75
     Saxons
    0.73
     myſelf
    0.73
    ſelves
    0.73
     يتيمه
    0.72
    Activations Density 0.003%

    No Known Activations