© 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. Gemma-2-2B
    5. 21-GEMMASCOPE-TRANSCODER-16K
    7. 3196
    Prev
    Next
    INDEX
    Explanations

    words related to how actions are done, as well as things that should not go together but do

    oai_token-act-pair · gemini-2.0-flash

    words or phrases about action and change

    np_token-act-pair-logits · gpt-4o-mini

    say "-ly" words

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2b-pt-transcoders/layer_21/width_16k/average_l0_13
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.21.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
    ValueStyle
    -0.61
    very
    -0.61
     very
    -0.60
    الإنجليزية
    -0.59
     صوتيه
    -0.57
    ंदीखरीदारी
    -0.56
     pratiquement
    -0.56
    quite
    -0.56
    muito
    -0.55
    RectangleBorder
    -0.54
    POSITIVE LOGITS
    ████
    0.54
     evacuated
    0.53
    kmale
    0.51
    hdashline
    0.51
    XMLSchema
    0.50
     confessed
    0.50
     misinterpreted
    0.49
     apologized
    0.49
     rent
    0.49
     mastered
    0.49
    Activations Density 4.693%

    No Known Activations