© 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
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 65248
    Prev
    Next
    INDEX
    Explanations

    prediction and outcomes

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

    Neuron 4 is tuned to words that signal outcomes or predictions—especially in sports or betting contexts (e.g. wins, falls, cover, close, outright).

    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
     was
    -1.11
     prezen
    -0.94
     También
    -0.92
    újo
    -0.91
    erschied
    -0.90
     other
    -0.89
    Эк
    -0.88
    Қ
    -0.85
    tbody
    -0.85
     guerres
    -0.85
    POSITIVE LOGITS
     will
    1.36
     should
    1.08
     expect
    1.05
     win
    1.02
     unless
    1.02
     predicts
    1.00
     prediction
    1.00
     shouldn
    0.98
    ētu
    0.97
     però
    0.96
    Activations Density 0.003%

    No Known Activations