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. 36-GEMMASCOPE-RES-16K
    6. 16307
    Prev
    Next
    INDEX
    Explanations

    legal terminology related to court decisions and legislative actions

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B @ 36-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-pt-res/layer_36/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.36.hook_resid_post
    Hook Layer
    36
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     muhte
    -0.56
     belki
    -0.49
     mükemmel
    -0.49
     unexpected
    -0.45
     hır
    -0.41
     probably
    -0.41
     harika
    -0.40
     inesperado
    -0.40
    unexpected
    -0.40
     vertraut
    -0.39
    POSITIVE LOGITS
    Certain
    0.82
     certain
    0.77
     Certain
    0.77
    certain
    0.75
     henceforth
    0.69
     anyone
    0.69
     Exceptions
    0.68
     visse
    0.66
     (§
    0.66
     unless
    0.66
    Activations Density 1.136%

    No Known Activations