© 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
    2. Gemma-2-27B
    3. 10-GEMMASCOPE-RES-131K
    4. 42714
    Prev
    Next
    INDEX
    Explanations

    growth factors

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

    XML and markup tags, particularly closing tags and numeric elements.

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

    The neuron fires on numeric tokens, especially multi‐digit numbers and decimals (e.g. version numbers, coordinates, measurements)

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
     било
    -1.58
    的特点
    -1.56
     dinosaurio
    -1.55
    when
    -1.55
    žiť
    -1.55
     lachen
    -1.51
    Поэтому
    -1.50
     какъ
    -1.49
     Detalles
    -1.49
     alebo
    -1.49
    POSITIVE LOGITS
     over
    1.73
     between
    1.63
     вернуться
    1.55
    都在
    1.52
    щита
    1.50
     on
    1.48
     everything
    1.48
    inerario
    1.47
     a
    1.45
    riebene
    1.44
    Activations Density 0.010%

    No Known Activations