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 - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 158
    Prev
    Next
    INDEX
    Explanations

    recurring phrases or the concept of repetition

    oai_token-act-pair · gpt-4o-miniTriggered by @bot

    phrases involving repetition or recurring events (especially the word "again").

    oai_token-act-pair · claude-3-5-haiku-20241022Triggered by @neilrathi
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B @ 20-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-pt-res/layer_20/width_16k/average_l0_68
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    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
     zwar
    -0.27
    ค
    -0.26
     Sécurité
    -0.25
     sometimes
    -0.23
    ước
    -0.23
     initial
    -0.23
     certain
    -0.23
    เบ
    -0.23
     quarante
    -0.22
     dům
    -0.22
    POSITIVE LOGITS
    again
    1.05
     again
    1.04
    Again
    1.02
     Again
    0.95
    今回も
    0.93
     AGAIN
    0.93
     nuevamente
    0.87
     novamente
    0.85
    __(/*!
    0.80
    AGAIN
    0.79
    Activations Density 0.307%

    No Known Activations