© 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. 28543
    Prev
    Next
    INDEX
    Explanations

    real report Experimental

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

    This neuron fires on content words in narrative prose (i.e. story-style descriptive text) rather than on code, metadata, or other non-narrative material.

    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
    PUESTA
    -1.43
    Inne
    -1.34
     Cataluña
    -1.32
    Komentar
    -1.28
     salonu
    -1.28
    bilden
    -1.27
    Fär
    -1.23
     architekt
    -1.23
     唯美
    -1.22
     Inglaterra
    -1.20
    POSITIVE LOGITS
    ৩
    1.30
     into
    1.26
     \
    1.26
    \
    1.23
    j
    1.21
     <
    1.20
    *
    1.19
    혀
    1.18
     /
    1.18
     |
    1.17
    Activations Density 0.004%

    No Known Activations