© 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. 124749
    Prev
    Next
    INDEX
    Explanations

    following text

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

    The neuron strongly activates on occurrences of the word “following” (and similar list‐or example‐introducing tokens) that introduce code or enumerated items.

    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
     possibility
    -1.53
     recomendado
    -1.33
     negras
    -1.13
    odrome
    -1.12
     mejores
    -1.11
     matricula
    -1.10
     can
    -1.10
     libras
    -1.09
     latter
    -1.07
    стій
    -1.06
    POSITIVE LOGITS
    Ikke
    1.43
    Etter
    1.33
     personer
    1.21
    猁
    1.13
    н
    1.13
     usługi
    1.12
    dass
    1.12
     musikk
    1.10
     韓国
    1.09
     őket
    1.08
    Activations Density 0.237%

    No Known Activations