© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW 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. 5250
    Prev
    Next
    INDEX
    Explanations

    programming or coding constructs and their related components

    oai_token-act-pair · gpt-4o-mini
    Triggered by @bot

    proper nouns and entities

    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-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
    Negative Logits
     Claim
    -0.44
     claim
    -0.43
     Kur
    -0.41
    Claim
    -0.38
     Inner
    -0.38
     Off
    -0.38
    Eig
    -0.37
     Efforts
    -0.37
     Claims
    -0.37
     inverse
    -0.37
    POSITIVE LOGITS
     Vordergrund
    0.54
     pleaſure
    0.53
     ProtoMessage
    0.52
     christlichen
    0.51
     enfans
    0.51
     sukienka
    0.51
     perfons
    0.50
     cucchiaio
    0.50
     indígen
    0.50
    üstü
    0.50
    Activations Density 0.127%

    No Known Activations