© 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-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 68754
    Prev
    Next
    INDEX
    Explanations

    words and codes

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

    The neuron strongly lights up on very short (mostly two‐character) tokens—i.e. it’s detecting little two-letter “words” or abbreviations.

    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
     didn
    -1.78
     only
    -1.71
     just
    -1.70
     started
    -1.60
     went
    -1.60
    というわけで
    -1.58
     promotes
    -1.58
     eliminates
    -1.55
     probably
    -1.48
     enhances
    -1.47
    POSITIVE LOGITS
    thday
    1.96
    ofday
    1.73
    ⚌
    1.64
    quely
    1.63
     Dinas
    1.61
    EVEN
    1.56
    andafter
    1.56
     trám
    1.56
    NNN
    1.54
    //{
    1.52
    Activations Density 0.004%

    No Known Activations