Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsBlogSlackPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-9B-IT
    4. Residual Stream - 131k
    5. 20-GEMMASCOPE-RES-131K
    6. 38933
    Prev
    Next
    INDEX
    Explanations

    instances of the word "nevertheless" or similar transitional phrases indicating contrast or concession

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k
    Configuration
    google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    131,072
    Data Type
    float32
    Hook Name
    blocks.20.hook_resid_post
    Hook Layer
    20
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    queryInterface
    -0.56
     ***!
    -0.56
    хьтан
    -0.54
     fure
    -0.51
    woofer
    -0.50
     snippetHide
    -0.50
    chaun
    -0.50
     InputDecoration
    -0.50
    HeaderComponent
    -0.50
    verwijspagina
    -0.48
    POSITIVE LOGITS
    Nevertheless
    0.93
     Nevertheless
    0.93
    Nonetheless
    0.87
     nevertheless
    0.86
     Nonetheless
    0.85
     nonetheless
    0.81
    theless
    0.81
    anmoins
    0.79
    etheless
    0.77
     Trotzdem
    0.75
    Activations Density 0.016%

    No Known Activations