© 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. 101648
    Prev
    Next
    INDEX
    Explanations

    looks followed by descriptions

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

    The neuron fires on positive evaluative words and phrases—especially praise or “good‐performance” descriptors (e.g. looks good, played well, positive).

    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
     aqueles
    -1.47
     rescate
    -1.45
     estabiliz
    -1.43
     kupa
    -1.42
     moldura
    -1.42
     asientos
    -1.38
     perfeitamente
    -1.37
     pulseira
    -1.36
     stary
    -1.34
    apka
    -1.32
    POSITIVE LOGITS
     in
    1.72
    ’
    1.68
     for
    1.55
    an
    1.41
    ка
    1.31
    ta
    1.30
    gen
    1.30
     as
    1.30
    ha
    1.28
    </em>
    1.24
    Activations Density 0.023%

    No Known Activations