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. 994
    Prev
    Next
    INDEX
    Explanations

    punctuation and special character usage

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

    punctuation marks and formatting symbols typically used in code or technical writing

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

    mathematical citations

    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
    AddTagHelper
    -0.87
    ⓧ
    -0.86
     समीक्षाओं
    -0.85
    AccessorTable
    -0.77
     @"/
    -0.75
    Lähteet
    -0.74
    Искәрмәләр
    -0.74
     gynhyrchwyd
    -0.72
    ########.
    -0.72
    webElementXpaths
    -0.70
    POSITIVE LOGITS
     fraî
    0.61
     étrangères
    0.60
     culturelles
    0.60
     grantor
    0.58
     kvinna
    0.55
     conquête
    0.55
     habet
    0.54
     oameni
    0.53
     rapides
    0.52
     féminine
    0.52
    Activations Density 0.003%

    No Known Activations