© 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
    3. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    5. Gemma-2-9B
    7. Residual Stream - 16k
    9. 3-GEMMASCOPE-RES-16K
    11. 8292
    Prev
    Next
    INDEX
    Explanations

    instances of the word "held" related to events or meetings

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B @ 3-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-pt-res/layer_3/width_16k/average_l0_90
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.3.hook_resid_post
    Hook Layer
    3
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    moon
    -0.52
    mora
    -0.49
     moon
    -0.46
     Stringer
    -0.46
    Moon
    -0.46
    Monitor
    -0.45
    usercontent
    -0.45
     translateY
    -0.44
     siena
    -0.44
     mum
    -0.43
    POSITIVE LOGITS
     held
    0.92
    held
    0.78
     HELD
    0.78
    Held
    0.77
     Held
    0.72
     presented
    0.64
     PRESENTED
    0.58
     Popular
    0.57
     Presented
    0.55
     POPULAR
    0.54
    Activations Density 0.057%

    No Known Activations