Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    INDEX
    Explanations

    loss

    np_max-act · gemini-2.0-flash

    words and phrases indicating defeat, losing, or trailing behind in competitive contexts.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @faiazazmain99

    descriptions of unfavorable status or outcomes—being behind, losing, or otherwise negative—often in competitive or evaluative contexts.

    oai_token-act-pair · gpt-5Triggered by @faiazazmain99
    New Auto-Interp
    Negative Logits
    æį¨
    -0.27
    ASK
    -0.25
    風
    -0.25
     лиÑĪ
    -0.25
     Winds
    -0.25
     wind
    -0.24
    .Resource
    -0.24
    ãĤ»ãĥ³ãĤ¿ãĥ¼
    -0.23
    VERAGE
    -0.23
     Amb
    -0.23
    POSITIVE LOGITS
    åħ±äº§
    0.27
    dee
    0.27
    adol
    0.26
     deepest
    0.26
    âĢ¦”
    0.26
    å¾Īæ·±
    0.26
    otel
    0.25
    emia
    0.25
    ULO
    0.24
    /popper
    0.24
    Act Density 27.446%

    No Known Activations