© 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. 22-GEMMASCOPE-TRANSCODER-16K
    7. 9658
    Prev
    Next
    INDEX
    Explanations

    hedging language using words like "may", "might", "would", or "must."

    oai_token-act-pair · gemini-2.0-flash

    say "have"

    np_max-act-logits · gemini-2.0-flash
    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
    were
    -0.67
     poz
    -0.52
     were
    -0.50
    ELER
    -0.49
     dis
    -0.49
     did
    -0.48
    Do
    -0.47
    즈
    -0.47
    was
    -0.47
    do
    -0.46
    POSITIVE LOGITS
     have
    0.90
     Have
    0.80
     hae
    0.78
     للاسماء
    0.77
     BorderRadius
    0.76
     itſelf
    0.75
     iſt
    0.74
    Have
    0.74
     hebben
    0.73
     bave
    0.72
    Activations Density 0.768%

    No Known Activations