© 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. 9543
    Prev
    Next
    INDEX
    Explanations

    where phrases or clauses

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

    The neuron is primarily triggering on the occurrences of the word “where” (and the immediately surrounding copular construction “where the … is”).

    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
    谥
    -1.09
     אנשים
    -1.05
     aceste
    -1.04
     фести
    -1.03
     faciliter
    -1.01
    惊艳
    -0.97
    冖
    -0.97
    这个问题
    -0.97
     touché
    -0.96
     peo
    -0.95
    POSITIVE LOGITS
     we
    1.30
     now
    1.29
     include
    1.13
     again
    1.10
     siento
    1.07
     because
    1.04
     Vereinigte
    1.01
     !!
    1.00
     nuevamente
    0.98
    because
    0.98
    Activations Density 0.005%

    No Known Activations