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. 16-GEMMASCOPE-RES-65K
    6. 53177
    Prev
    Next
    INDEX
    Explanations

    references to food substitutes and their ingredients

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

    simulated or artificial

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 16-gemmascope-res-65k
    Configuration
    google/gemma-scope-2b-pt-res/layer_16/width_65k/average_l0_128
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    65,536
    Data Type
    float32
    Hook Name
    blocks.16.hook_resid_post
    Hook Layer
    16
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    itespace
    -0.57
     sementes
    -0.52
    concel
    -0.51
    ModelAdmin
    -0.49
    Tikang
    -0.48
    หวัด
    -0.48
     poésie
    -0.48
    uppercase
    -0.47
    cipline
    -0.46
    tune
    -0.46
    POSITIVE LOGITS
     realism
    1.23
     simulating
    1.20
     simulates
    1.19
     simulate
    1.19
     simulated
    1.17
     fake
    1.15
     realistic
    1.15
     realista
    1.13
     réaliste
    1.09
     Realistic
    1.09
    Activations Density 0.264%

    No Known Activations