© 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
    3. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    5. Gemma-2-9B
    7. Residual Stream - 16k
    9. 35-GEMMASCOPE-RES-16K
    11. 7147
    Prev
    Next
    INDEX
    Explanations

    instances of mistakes or errors in various contexts and their implications

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B @ 35-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-pt-res/layer_35/width_16k/average_l0_120
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.35.hook_resid_post
    Hook Layer
    35
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    Diweddarwch
    -0.44
    拋
    -0.39
    PLAUSE
    -0.38
    httphttps
    -0.37
    átky
    -0.36
     contextLoads
    -0.35
    ennom
    -0.35
     Numerade
    -0.34
     UserDao
    -0.34
    ernalia
    -0.34
    POSITIVE LOGITS
     mistake
    2.58
     mistakes
    2.27
     mistakenly
    2.06
     mistaken
    2.06
     Mistake
    2.03
     Mistakes
    1.94
     error
    1.91
    Mistake
    1.90
     errors
    1.82
     erreur
    1.79
    Activations Density 1.616%

    No Known Activations