© 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
    3. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    5. Gemma-2-9B
    7. Residual Stream - 131k
    9. 20-GEMMASCOPE-RES-131K
    11. 104089
    Prev
    Next
    INDEX
    Explanations

    words that indicate importance or significance in context

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

    Words that serve as textual pivots or anchors at phrase endings, often indicating conclusion of an idea, transition between thoughts, or reference to previously established information.

    eleuther_acts_top20 · claude-3-7-sonnet-20250219Triggered by @lu-christina
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-9b-pt-res/layer_20/width_131k/average_l0_114
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    131,072
    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
    "]));
    -0.56
     besseren
    -0.56
    )++;
    -0.56
     leeren
    -0.55
    "}")
    -0.55
    IsContent
    -0.54
    ")));
    -0.54
    Geplaatst
    -0.54
     Wikiseite
    -0.54
    })(
    -0.52
    POSITIVE LOGITS
    SequentialGroup
    0.44
     <<<<<<<<<<<<<<
    0.34
     rè
    0.33
    IBarButtonItem
    0.32
     ModelRenderer
    0.30
    withIdentifier
    0.28
     inds
    0.28
     Archer
    0.28
    pAd
    0.28
     Davis
    0.28
    Activations Density 0.088%

    No Known Activations