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

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI 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. 60654
    Prev
    Next
    INDEX
    Explanations

    discussions about societal issues and moral dilemmas

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

    harmful topics

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

    the neuron detects racist or strongly derogatory language aimed at social groups (demeaning/offensive statements).

    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
    ကိုးကား
    -0.44
     bune
    -0.39
    +:+
    -0.36
    赦
    -0.34
    ภาค
    -0.34
     heureux
    -0.34
     excelencia
    -0.34
    onViewCreated
    -0.33
    orsche
    -0.32
    eterangan
    -0.31
    POSITIVE LOGITS
     NSCoder
    0.60
    queryInterface
    0.52
     prostitution
    0.50
    ThroughAttribute
    0.49
     illegal
    0.49
     onPostExecute
    0.49
    DebuggerNonUser
    0.49
     harmful
    0.48
     behaviors
    0.47
    Bibliograf
    0.47
    Activations Density 0.096%

    No Known Activations