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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-9B
    4. Residual Stream - 131k
    5. 23-GEMMASCOPE-RES-131K
    6. 107
    Prev
    Next
    INDEX
    Explanations

    instances of the word "earlier."

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-9b-pt-res/layer_23/width_131k/average_l0_101
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    131,072
    Data Type
    float32
    Hook Name
    blocks.23.hook_resid_post
    Hook Layer
    23
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    TypedDataSet
    -0.57
    DataAnnotations
    -0.52
    moduleName
    -0.50
     candle
    -0.47
     Sulf
    -0.46
     Sting
    -0.46
     createState
    -0.45
     HYDRO
    -0.44
     Scientology
    -0.43
     Knockout
    -0.43
    POSITIVE LOGITS
    earlier
    1.04
     earlier
    1.00
     Earlier
    0.91
    Earlier
    0.89
     eerder
    0.64
    previous
    0.63
     wcześ
    0.61
     früheren
    0.60
    previously
    0.58
     previous
    0.57
    Activations Density 0.009%

    No Known Activations