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-2-9B
    4. Residual Stream - 16k
    5. 11-GEMMASCOPE-RES-16K
    6. 5189
    Prev
    Next
    INDEX
    Explanations

    instances of the word "press" and its variations

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B @ 11-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-pt-res/layer_11/width_16k/average_l0_118
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.11.hook_resid_post
    Hook Layer
    11
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ftagPool
    -0.51
    LLocation
    -0.50
    RectangleBorder
    -0.48
    signable
    -0.47
    OrNil
    -0.47
    enumii
    -0.46
     дописавши
    -0.44
     onOptions
    -0.43
     Quaternion
    -0.42
     ProtoMessage
    -0.41
    POSITIVE LOGITS
     Press
    1.66
     press
    1.60
     PRESS
    1.52
    Press
    1.49
    press
    1.47
     presses
    1.34
     Pressing
    1.30
    PRESS
    1.29
     pressed
    1.26
     Presses
    1.21
    Activations Density 0.122%

    No Known Activations