© 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
    2. Gemma-2-27B
    3. 22-GEMMASCOPE-RES-131K
    4. 54074
    Prev
    Next
    INDEX
    Explanations

    end of sentence

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

    The neuron strongly activates on the very first content word of a new sentence or quoted line—that is, it marks sentence or dialogue-segment beginnings.

    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
     modrá
    -1.24
     páns
    -1.20
    ducir
    -1.18
     dáms
    -1.13
     červená
    -1.10
     menyiapkan
    -1.09
     comenzado
    -1.05
     dámské
    -1.04
     atractivo
    -1.02
     efficacité
    -1.02
    POSITIVE LOGITS
     he
    1.23
     then
    1.16
     was
    1.13
     because
    1.11
     would
    1.09
     later
    1.09
     from
    1.08
     so
    1.04
    라는
    1.02
     panna
    1.01
    Activations Density 0.010%

    No Known Activations