© 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-3-27B-IT
    5. 31-GEMMASCOPE-2-RES-65K
    7. 51526
    Prev
    Next
    INDEX
    Explanations

    phrases expressing humor or comedic elements related to exhaustion or tiredness.

    oai_token-act-pair · gpt-4o-miniTriggered by @gersonkroiz

    count and counterfactual

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-27b-it/resid_post/layer_31_width_65k_l0_medium
    Prompts (Dashboard)
    238,145 prompts, 512 tokens each
    Dataset (Dashboard)
    lmsys + oasst1
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    س
    1.24
    स
    0.86
    я
    0.80
    ك
    0.75
    સ
    0.75
    ক
    0.73
    ী
    0.73
    ை
    0.72
    ल
    0.70
    ق
    0.69
    POSITIVE LOGITS
    enance
    0.89
    る
    0.80
    w
    0.79
    count
    0.72
    counter
    0.72
    r
    0.72
    n
    0.68
    west
    0.68
    eren
    0.67
    for
    0.66
    Activations Density 0.035%

    No Known Activations