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. 28532
    Prev
    Next
    INDEX
    Explanations

    expressions of fear and cognitive difficulties in hypothetical scenarios

    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
    ⓧ
    -0.78
     butterknife
    -0.65
    PyErr
    -0.61
     dotenv
    -0.56
    πουργ
    -0.54
    ∬
    -0.53
     Numerology
    -0.53
     Suits
    -0.53
    EndTag
    -0.52
    ámb
    -0.52
    POSITIVE LOGITS
     ostavi
    0.61
    FailureListener
    0.60
    自分は
    0.60
     EconPapers
    0.59
    =?";
    0.59
    AlterField
    0.57
     somehow
    0.56
    InstrumentedTest
    0.54
     المعيارى
    0.54
     للاسماء
    0.53
    Activations Density 0.150%

    No Known Activations