© 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
    2. Gemma-2-27B
    3. 10-GEMMASCOPE-RES-131K
    4. 128747
    Prev
    Next
    INDEX
    Explanations

    identifying attribute names

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

    This neuron fires on attribute‐name tokens (and their accompanying equals signs) within XML/markup tags.

    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
     where
    -1.59
     with
    -1.51
     at
    -1.50
     in
    -1.49
     when
    -1.44
     chose
    -1.44
     gave
    -1.41
     and
    -1.35
     who
    -1.33
     as
    -1.29
    POSITIVE LOGITS
    탉
    1.50
     gênero
    1.47
     jeste
    1.47
    킁
    1.41
     così
    1.39
    
    1.39
     território
    1.38
     princípio
    1.36
     предоставля
    1.34
    --"
    1.33
    Activations Density 0.069%

    No Known Activations