© 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. 54011
    Prev
    Next
    INDEX
    Explanations

    problems with foreign words

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

    The neuron fires on sentences that describe a change in behavior or user‐impact (e.g. “This was working… but… it no longer does” or “Will the users find it annoying…?”).

    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
    robat
    -1.33
     AUSSI
    -1.27
     նաև
    -1.18
    時も
    -1.17
    何より
    -1.17
    を高
    -1.14
    澌
    -1.14
    リラックス
    -1.14
    ceiver
    -1.13
    igerung
    -1.13
    POSITIVE LOGITS
    ต้อง
    1.45
     problems
    1.32
     would
    1.29
     unable
    1.27
    จะ
    1.23
    問題
    1.20
    InjectAttribute
    1.20
     Nước
    1.18
    问题
    1.18
     समस्या
    1.16
    Activations Density 0.082%

    No Known Activations