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

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    5. Gemma-2-9B
    7. Residual Stream - 16k
    9. 1-GEMMASCOPE-RES-16K
    11. 10661
    Prev
    Next
    INDEX
    Explanations

    instances of the word "worse" or its variations, indicating degradation or negative comparisons

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B @ 1-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-pt-res/layer_1/width_16k/average_l0_69
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.1.hook_resid_post
    Hook Layer
    1
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Biôgrafia
    -0.55
    Lass
    -0.50
    bulin
    -0.48
     Lass
    -0.47
     INSEE
    -0.47
     HasFactory
    -0.47
    ikistan
    -0.45
     Bulk
    -0.45
    audiovisuel
    -0.45
     Sis
    -0.45
    POSITIVE LOGITS
    __":
    0.60
    __':
    0.59
    __':
    0.58
    __":
    0.57
     worse
    0.57
    setFont
    0.56
     Tre
    0.55
    Hochspringen
    0.54
     worst
    0.53
    AppCompat
    0.52
    Activations Density 0.305%

    No Known Activations