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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. GPT2-Small
    3. Transcoders Residuals
    4. 8-TRES-DC
    5. 400
    Prev
    Next
    INDEX
    Explanations

    the conjunction "but" used to introduce contrasting statements

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    learn
    -0.59
     (−
    -0.59
    miss
    -0.58
    き
    -0.55
    DEBUG
    -0.54
    hyper
    -0.53
    hof
    -0.50
    erk
    -0.49
    exit
    -0.49
    hom
    -0.49
    POSITIVE LOGITS
     interestingly
    0.70
     secondly
    0.68
     luckily
    0.67
     hey
    0.65
     unfortunately
    0.62
     alas
    0.61
     importantly
    0.59
     sadly
    0.59
     thankfully
    0.59
     fortunately
    0.56
    Activations Density 0.087%

    No Known Activations