© 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
    4. Residual Stream - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 8948
    Prev
    Next
    INDEX
    Explanations

    elements related to bug reporting and troubleshooting processes

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

    language associated with software issue tracking and QA workflows, including bug statuses, update requests, and steps to reproduce or fix reported problems.

    oai_token-act-pair · gpt-5Triggered by @jyhe0408

    text related to bug reports, issue tracking, and responses to requests or complaints.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B @ 20-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-pt-res/layer_20/width_16k/average_l0_68
    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
     للاسماء
    -0.97
     الحره
    -0.61
    VIAF
    -0.60
    
    -0.58
    jsxFileName
    -0.57
    EnableWeb
    -0.55
    enderror
    -0.54
    󠁮
    -0.54
    :✨
    -0.52
    ſt
    -0.52
    POSITIVE LOGITS
     investigate
    0.43
     respond
    0.42
     response
    0.39
     investigation
    0.38
     promise
    0.38
     responded
    0.38
     responds
    0.38
     reform
    0.38
     Response
    0.37
     investigates
    0.36
    Activations Density 0.259%

    No Known Activations