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

    Neuronpedia

    INDEX
    Explanations

    PodsDummy

    np_max-act · gemini-2.0-flash

    tokens that look like code or technical identifiers (CamelCase/PascalCase names, escaped sequences, usernames, and other non-natural-language tokens).

    oai_token-act-pair · gpt-5-miniTriggered by @vetterc0
    New Auto-Interp
    Negative Logits
    0
    -0.08
    エ
    -0.07
                                 ↵
    -0.06
    Son
    -0.06
    OTT
    -0.06
    	        ↵
    -0.06
           ↵↵
    -0.06
    क
    -0.06
                                                    ↵
    -0.06
    ではない
    -0.06
    POSITIVE LOGITS
    Digest
    0.07
    _combo
    0.07
    (Dense
    0.06
    (pair
    0.06
    NSMutable
    0.06
     ук
    0.06
     Liver
    0.06
     violin
    0.06
    _episodes
    0.06
    _again
    0.06
    Act Density 0.242%

    No Known Activations