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. 24-GEMMASCOPE-RES-131K
    6. 764
    Prev
    Next
    INDEX
    Explanations

    punctuation marks, particularly commas, in various contexts

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

    transition words then new

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-9b-pt-res/layer_24/width_131k/average_l0_97
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    131,072
    Data Type
    float32
    Hook Name
    blocks.24.hook_resid_post
    Hook Layer
    24
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     TS
    -0.50
     ')
    
    -0.50
    EndGlobal
    -0.49
     Chriftian
    -0.48
    baskets
    -0.48
    desertcart
    -0.48
     )))
    -0.48
    QRST
    -0.48
    bag
    -0.48
     Genie
    -0.48
    POSITIVE LOGITS
    addPreferredGap
    0.52
     informée
    0.42
    ,
    0.40
    Within
    0.38
    Décès
    0.38
    Legături
    0.38
    addGap
    0.38
     within
    0.38
     pēc
    0.37
    verifyException
    0.37
    Activations Density 0.029%

    No Known Activations