© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 9969
    Prev
    Next
    INDEX
    Explanations

    phoneme, codon, epitope

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

    linguistics terms, particularly phonemes and their related terms (like epitope, codon) in technical or academic texts.

    oai_token-act-pair · claude-3-7-sonnet-20250219Triggered by @neilrathi

    This neuron strongly activates on the “-me” portion of “phoneme” (and on the split tokens “phone” + “me”), i.e. it is detecting the “me” suffix in words like “phoneme.”

    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
     сист
    -2.09
     Like
    -1.91
    芣
    -1.91
     одном
    -1.88
    ⌇
    -1.84
    ”？
    -1.80
    鐿
    -1.80
    ”-
    -1.76
    “-
    -1.76
     They
    -1.74
    POSITIVE LOGITS
     金属
    1.81
    凄く
    1.77
     they
    1.73
      
    1.69
    bisous
    1.68
     küçük
    1.66
    '
    1.65
     devido
    1.63
     that
    1.63
    łóż
    1.57
    Activations Density 0.029%

    No Known Activations