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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & 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. 18-GEMMASCOPE-RES-16K
    6. 8528
    Prev
    Next
    INDEX
    Explanations

    elements relating to mathematical or statistical concepts

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

    Conditional words (if, when, unless)

    np_acts-logits-general · gemini-2.0-flash

    conditional or causal conjunctions

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 18-gemmascope-res-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_18/width_16k/average_l0_74
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.18.hook_resid_post
    Hook Layer
    18
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    AsUp
    -0.71
    httphttps
    -0.71
     lenker
    -0.63
     TextFormField
    -0.61
     Monfieur
    -0.60
     ―――――
    -0.59
     useDispatch
    -0.58
    naut
    -0.58
     Reſ
    -0.57
    createCanvas
    -0.57
    POSITIVE LOGITS
     if
    1.38
     when
    1.21
     wenn
    1.01
     если
    1.01
     якщо
    1.01
    if
    0.99
     assuming
    0.95
    when
    0.94
     because
    0.92
     gdy
    0.89
    Activations Density 0.299%

    No Known Activations