© 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. 34-GEMMASCOPE-RES-131K
    7. 92226
    Prev
    Next
    INDEX
    Explanations

    George Will OR salinity challenge

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

    The neuron fires on numeric tokens—especially longer or out‐of‐context digit sequences (e.g. large standalone numbers often found in code, diffs, or timestamps).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
     what
    -1.38
     unique
    -1.25
     our
    -1.17
     your
    -1.16
     both
    -1.13
     this
    -1.09
    ınıza
    -1.08
    تاح
    -1.05
     hacían
    -1.03
     triko
    -1.02
    POSITIVE LOGITS
    mathrm
    1.10
    在这种
    1.05
     etern
    0.98
     メル
    0.96
    段时间
    0.94
    所以在
    0.94
     McDonalds
    0.91
     レン
    0.91
     commando
    0.91
     ........
    0.91
    Activations Density 0.000%

    No Known Activations