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. 981
    Prev
    Next
    INDEX
    Explanations

    instances of numerical values and their associated formats

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

    copyright and license-related information

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

    software copyright notices

    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
    bootstrapcdn
    -0.65
     nahilalakip
    -0.57
    তথ্যসূত্র
    -0.55
    ")));
    
    -0.54
     statt
    -0.54
    )‏
    -0.53
    mybatisplus
    -0.52
    Халык
    -0.52
    Referências
    -0.51
    Personensuche
    -0.50
    POSITIVE LOGITS
    Portale
    0.76
     HasFactory
    0.63
    Bibliograf
    0.58
    tetten
    0.53
     وتسجيلات
    0.53
     MainAxisSize
    0.51
    клопе
    0.51
     noix
    0.49
    évaluateur
    0.49
     xmlns
    0.49
    Activations Density 0.018%

    No Known Activations