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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-9B
    4. Residual Stream - 16k
    5. 37-GEMMASCOPE-RES-16K
    6. 1166
    Prev
    Next
    INDEX
    Explanations

    expressions and variations of the word "love."

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B @ 37-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-pt-res/layer_37/width_16k/average_l0_124
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.37.hook_resid_post
    Hook Layer
    37
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Pert
    -0.33
     Tại
    -0.33
    öffnung
    -0.30
    XHR
    -0.30
     كومونز
    -0.29
    Cpf
    -0.29
    őzés
    -0.29
    至於
    -0.29
     Proud
    -0.29
    ところで
    -0.28
    POSITIVE LOGITS
    hate
    0.76
    liest
    0.75
    Hate
    0.73
     affair
    0.70
    birds
    0.70
    letters
    0.69
     hate
    0.64
     letters
    0.64
     Hate
    0.63
    bug
    0.60
    Activations Density 0.062%

    No Known Activations