© 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. 34-GEMMASCOPE-RES-131K
    7. 33741
    Prev
    Next
    INDEX
    Explanations

    changing score context

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

    The neuron activates on words describing a team’s scoring burst or comeback (e.g. “rallied,” numeric run counts like “five,” inning/period markers, “scored,” “run to take the lead,” etc.).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
     atrak
    -0.85
    ᐳ
    -0.80
     Cardiovascular
    -0.79
    klare
    -0.77
     seksual
    -0.76
     rechazo
    -0.76
     precisión
    -0.75
     eksper
    -0.75
    gangss
    -0.75
    anair
    -0.74
    POSITIVE LOGITS
     leveling
    3.19
     equal
    3.14
     levelling
    2.88
     equalizer
    2.88
     leveled
    2.77
     levelled
    2.73
     equalization
    2.66
     Equal
    2.52
    equal
    2.47
    Equal
    2.45
    Activations Density 0.018%

    No Known Activations