© 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. Gemma-2-2B
    5. 12-RES-MATRYOSHKA-DC
    7. 5216
    Prev
    Next
    INDEX
    Explanations

    instances of the word "hold" in various contexts

    oai_token-act-pair · gpt-4o-mini
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    chanind/gemma-2-2b-batch-topk-matryoshka-saes-w-32k-l0-40/standard/blocks.12.hook_resid_post
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    32,768
    Data Type
    float32
    Hook Name
    blocks.12.hook_resid_post
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    chanind/pile-uncopyrighted-gemma-1024-abbrv-1B
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    PyExc
    -0.75
     bác
    -0.74
    genossen
    -0.74
     routes
    -0.70
     jSON
    -0.70
    EEN
    -0.69
     незавершена
    -0.69
    Authors
    -0.69
     lèvres
    -0.68
    getDescription
    -0.68
    POSITIVE LOGITS
     Hold
    1.48
     hold
    1.46
     holds
    1.46
    holds
    1.46
     Holds
    1.43
     HOLD
    1.40
     holding
    1.39
     Holding
    1.39
    Hold
    1.38
     Holden
    1.36
    Activations Density 0.200%

    No Known Activations