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

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    5. Gemma-2-9B
    7. Residual Stream - 16k
    9. 0-GEMMASCOPE-RES-16K
    11. 5837
    Prev
    Next
    INDEX
    Explanations

    instances of the word "proper."

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B @ 0-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-pt-res/layer_0/width_16k/average_l0_129
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.0.hook_resid_post
    Hook Layer
    0
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    __()
    -0.46
     fuck
    -0.44
    catraz
    -0.43
    revenge
    -0.43
     fucked
    -0.42
    stealth
    -0.42
     smir
    -0.41
     zip
    -0.41
     BBQ
    -0.40
    </thead>
    -0.40
    POSITIVE LOGITS
     Proper
    0.93
    Proper
    0.92
     proper
    0.90
     improper
    0.71
    proper
    0.71
     PROPER
    0.65
     Improper
    0.64
     Sister
    0.58
     Properly
    0.56
     properly
    0.55
    Activations Density 0.291%

    No Known Activations