© 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. 10-GEMMASCOPE-RES-131K
    7. 105617
    Prev
    Next
    INDEX
    Explanations

    acquired, funded, driven, operated

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

    The neuron consistently fires on hyphen/dash tokens (especially when they join words or appear with numbers), i.e. it’s detecting the presence of hyphens.

    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
     minimale
    -1.58
    lieving
    -1.55
     sekarang
    -1.54
    骓
    -1.50
    ᕱ
    -1.49
    РЫ
    -1.49
    𓆸
    -1.48
    IEDAD
    -1.39
     kompeti
    -1.38
    ньому
    -1.38
    POSITIVE LOGITS
    ↵↵
    1.74
    c
    1.51
     which
    1.51
    ↵↵↵
    1.47
    antit
    1.44
    /
    1.44
     inspired
    1.36
     approach
    1.32
    f
    1.31
    anmoins
    1.29
    Activations Density 0.152%

    No Known Activations