© 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. 71523
    Prev
    Next
    INDEX
    Explanations

    contractions like didn't

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

    The neuron is essentially detecting very common high-frequency “stop” words (common function words) in the text.

    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
     rir
    -1.03
    Segue
    -0.97
     proce
    -0.95
     pne
    -0.94
     trase
    -0.91
    rindo
    -0.90
    Crowd
    -0.87
     zowel
    -0.86
     tampa
    -0.85
     bole
    -0.85
    POSITIVE LOGITS
     meer
    1.02
    GGLE
    0.89
     muchos
    0.89
     demandent
    0.88
     meses
    0.87
    nextProps
    0.87
     milliers
    0.86
     cruce
    0.85
    fR
    0.84
     deze
    0.82
    Activations Density 0.022%

    No Known Activations