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. 1096
    Prev
    Next
    INDEX
    Explanations

    references to quotes or sayings from notable figures

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

    punctuation marks and formatting elements in text

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

    famous philosophers and writers

    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
     ModelExpression
    -0.96
    WebElementEntity
    -0.93
    XmlAccessType
    -0.91
    ReusableCell
    -0.90
     المعيارى
    -0.88
     doInBackground
    -0.86
    
    -0.84
    tagHelperRunner
    -0.82
     CreateTagHelper
    -0.82
    CppMethod
    -0.82
    POSITIVE LOGITS
     unknown
    0.66
    Unknown
    0.65
     Unknown
    0.57
    unknown
    0.56
     prover
    0.51
     smer
    0.51
     unbekannt
    0.50
     coper
    0.47
     undetermined
    0.46
     kaž
    0.45
    Activations Density 0.031%

    No Known Activations