© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-2B
    5. 22-GEMMASCOPE-TRANSCODER-16K
    7. 4989
    Prev
    Next
    INDEX
    Explanations

    descriptions of violence and weapons, particularly knives and other makeshift weapons, and sometimes injuries.

    oai_token-act-pair · gemini-2.0-flash

    weapons

    np_max-act-logits · gemini-2.0-flash

    weapons

    np_max-act-logits · gpt-5-miniTriggered by @chenshw0109
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2b-pt-transcoders/layer_22/width_16k/average_l0_15
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.22.ln2.hook_normalized
    Architecture
    jumprelu_transcoder
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     die
    -0.50
     La
    -0.45
     asequ
    -0.43
     Die
    -0.42
     che
    -0.42
    varet
    -0.40
     élas
    -0.40
    活
    -0.40
     Un
    -0.40
     fiebre
    -0.40
    POSITIVE LOGITS
     uſed
    0.86
     itſelf
    0.83
     myſelf
    0.82
     Efq
    0.81
     himſelf
    0.79
     ostavi
    0.77
     againſt
    0.75
     whoſe
    0.75
     الحره
    0.74
    ItemBackground
    0.73
    Activations Density 10.027%

    No Known Activations