© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-9B-IT
    4. Residual Stream - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 3987
    Prev
    Next
    INDEX
    Explanations

    phrases indicating division and separation of items or subjects

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

    division or separation

    np_acts-logits-general · gemini-2.0-flash

    dividing into parts or between

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91
    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
    UserScript
    -0.57
    Données
    -0.50
    expandindo
    -0.49
    SuccessListener
    -0.49
    /**
    -0.49
     NavController
    -0.48
    Ссылки
    -0.47
    jspx
    -0.46
    IsPostBack
    -0.45
    lccn
    -0.45
    POSITIVE LOGITS
     split
    1.20
     Split
    1.02
    split
    1.01
     splitting
    1.00
     divide
    0.97
     splits
    0.94
    Split
    0.91
     divided
    0.90
     divides
    0.90
     separation
    0.89
    Activations Density 0.359%

    No Known Activations