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. 24-GEMMASCOPE-RES-16K
    6. 2275
    Prev
    Next
    INDEX
    Explanations

    terms related to mechanical operation or manipulation

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

    Adverbs describing gentle or slow actions

    np_acts-logits-general · gemini-2.0-flash

    slowly or gently performing actions

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 24-gemmascope-res-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_24/width_16k/average_l0_73
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.24.hook_resid_post
    Hook Layer
    24
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    WriteLiteral
    -0.64
     Programmable
    -0.60
     @"/
    -0.56
    WebServlet
    -0.55
     pungkasnya
    -0.53
     OnInit
    -0.50
     survives
    -0.50
     بيها
    -0.49
    ellino
    -0.49
    Kup
    -0.49
    POSITIVE LOGITS
     gently
    1.02
     slowly
    0.95
     Gently
    0.91
    Slowly
    0.90
     Slowly
    0.89
    slowly
    0.79
     GENT
    0.77
     langsam
    0.76
     SLOW
    0.75
     lightly
    0.74
    Activations Density 0.313%

    No Known Activations