© 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. 71721
    Prev
    Next
    INDEX
    Explanations

    software versions and build info

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

    The neuron strongly activates on digits, dots, slashes, and related punctuation that typically occur in version numbers or numeric metadata strings.

    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
     barnet
    -0.98
     devrez
    -0.96
     岡
    -0.96
     přij
    -0.92
     who
    -0.92
     pæ
    -0.92
     still
    -0.91
    oreille
    -0.91
     geluk
    -0.91
     artikelen
    -0.91
    POSITIVE LOGITS
    Boxscore
    1.20
    |}\
    1.03
     schloss
    1.00
     Ге
    0.96
    gbc
    0.94
     따라
    0.93
    mphony
    0.93
     ungefähr
    0.92
     argumentos
    0.92
    yled
    0.92
    Activations Density 0.006%

    No Known Activations