© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-9B
    4. Residual Stream - 16k
    5. 41-GEMMASCOPE-RES-16K
    6. 7457
    Prev
    Next
    INDEX
    Explanations

    phrases that question or emphasize the concept of "how."

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B @ 41-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-pt-res/layer_41/width_16k/average_l0_113
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.41.hook_resid_post
    Hook Layer
    41
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    Very
    -0.77
     Very
    -0.73
     Очень
    -0.55
    Muy
    -0.52
     rất
    -0.51
     Quite
    -0.49
    very
    -0.48
    homonymie
    -0.48
     Muy
    -0.45
     Très
    -0.44
    POSITIVE LOGITS
     much
    1.13
     badly
    0.83
     close
    0.80
     quickly
    0.78
     closely
    0.76
     important
    0.76
     difficult
    0.73
     powerful
    0.73
     strong
    0.73
     easy
    0.72
    Activations Density 0.300%

    No Known Activations