Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit 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 - 16k
    5. 4-GEMMASCOPE-RES-16K
    6. 11022
    Prev
    Next
    INDEX
    Explanations

    terms related to improvement or augmentation

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 4-gemmascope-res-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_4/width_16k/average_l0_124
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.4.hook_resid_post
    Hook Layer
    4
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     uParam
    -0.75
     Salad
    -0.72
    ậu
    -0.70
    oglo
    -0.69
    мені
    -0.68
    itoneum
    -0.64
    пуль
    -0.63
     salad
    -0.63
    Salad
    -0.62
    м
    -0.62
    POSITIVE LOGITS
     enhancement
    1.88
    hancing
    1.75
     enhancements
    1.74
    enhance
    1.73
     Enhance
    1.72
     enhance
    1.71
     Enhancement
    1.69
    Enhance
    1.68
     enhanced
    1.64
     enhancing
    1.64
    Activations Density 0.167%

    No Known Activations