© 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-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 16317
    Prev
    Next
    INDEX
    Explanations

    covalently bound, biomedical research

    np_acts-logits-general · gemini-2.5-flash-lite

    semicolons and punctuation especially in technical or coding contexts.

    oai_token-act-pair · claude-3-7-sonnet-20250219Triggered by @neilrathi

    The neuron is primarily detecting technical numeric literals and measurement‐style tokens (e.g. sizes, counts, boolean flags) in code or scientific text.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     comprenant
    -1.11
    感じです
    -1.01
     portent
    -0.96
    感じで
    -0.96
     chamado
    -0.91
     ersten
    -0.91
    のでしょうか
    -0.90
    ₒ
    -0.89
    ARRAY
    -0.89
     They
    -0.89
    POSITIVE LOGITS
     because
    1.19
     probably
    1.14
     maintain
    1.09
     occasionally
    1.06
     definitely
    1.05
     chose
    1.03
     virtually
    1.02
     nearing
    1.02
     more
    1.00
     choose
    1.00
    Activations Density 0.020%

    No Known Activations