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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-2B
    4. Residual Stream - 65k
    5. 12-GEMMASCOPE-RES-65K
    6. 13738
    Prev
    Next
    INDEX
    Explanations

    references to training or training programs

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 12-gemmascope-res-65k
    Configuration
    google/gemma-scope-2b-pt-res/layer_12/width_65k/average_l0_72
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    65,536
    Data Type
    float32
    Hook Name
    blocks.12.hook_resid_post
    Hook Layer
    12
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Antara
    -0.86
     KeyCode
    -0.84
     Osbourne
    -0.81
     Diocese
    -0.79
     écl
    -0.76
     Neel
    -0.75
     Edd
    -0.74
     Necklace
    -0.73
     Satellites
    -0.72
     Clarinet
    -0.71
    POSITIVE LOGITS
     training
    1.13
     trainings
    0.97
     Training
    0.95
    Training
    0.90
    training
    0.87
    TRAINING
    0.82
     trainers
    0.80
     TRAINING
    0.78
     Trainings
    0.78
     train
    0.75
    Activations Density 0.011%

    No Known Activations