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. 33-GEMMASCOPE-RES-131K
    6. 65924
    Prev
    Next
    INDEX
    Explanations

    punctuation marks and conversational cues

    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_33/width_131k/average_l0_91
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    131,072
    Data Type
    float32
    Hook Name
    blocks.33.hook_resid_post
    Hook Layer
    33
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     literally
    -0.29
     versus
    -0.29
    格好
    -0.29
     meanwhile
    -0.28
     contempor
    -0.28
     rather
    -0.28
    而不是
    -0.28
    拥
    -0.27
    timeScale
    -0.27
     dirait
    -0.27
    POSITIVE LOGITS
     wink
    0.80
    wink
    0.79
    MigrationBuilder
    0.70
    cough
    0.67
    Ahem
    0.66
    🤫
    0.65
     shameless
    0.65
     hush
    0.63
     winked
    0.63
    ask
    0.60
    Activations Density 0.036%

    No Known Activations