© 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-IT
    7. Residual Stream - 131k
    9. 20-GEMMASCOPE-RES-131K
    11. 98064
    Prev
    Next
    INDEX
    Explanations

    terms related to digital media and content

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

    science, contracts, travel, or user

    np_acts-logits-general · gemini-2.5-flash-lite

    The neuron detects very high-frequency common English words and discourse markers (i.e., common function/content tokens used for general reference or example).

    oai_token-act-pair · gpt-5-miniTriggered by @vetterc0
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k
    Configuration
    google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81
    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
    djangoproject
    -0.40
     referrerpolicy
    -0.40
    wnętr
    -0.40
     lämp
    -0.36
    ypeł
    -0.36
     hiyo
    -0.36
     Mischung
    -0.35
     ndani
    -0.35
     jantung
    -0.35
     Spitze
    -0.35
    POSITIVE LOGITS
     traditions
    0.57
    ஦
    0.55
    LookAnd
    0.55
    ">//
    0.54
    <unused51>
    0.54
    <unused41>
    0.54
    <unused74>
    0.54
     natal
    0.54
    [@BOS@]
    0.54
    <unused23>
    0.54
    Activations Density 0.054%

    No Known Activations