Explanations

    phrases indicating falsehood or deception

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    Juliushanhanhan/llama-3-8b-it-res/blocks.25.hook_resid_post
    Features
    65,536
    Data Type
    float32
    Hook Name
    blocks.25.hook_resid_post
    Hook Layer
    25
    Architecture
    gated
    Context Size
    1,024
    Dataset
    Juliushanhanhan/openwebtext-1b-llama3-tokenized-cxt-1024
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    phrases indicating falsehood or deception
    Neuronpedia logo
    LLAMA3-8B-IT
    25-RES-JH
    INDEX 23610
    Negative Logits
    luk
    -0.16
    mlin
    -0.15
    CHandle
    -0.15
    opic
    -0.15
    輪
    -0.15
    é¡Į
    -0.14
    _Null
    -0.14
    avor
    -0.14
    etz
    -0.14
    alnız
    -0.14
    POSITIVE LOGITS
     lie
    0.97
     lies
    0.94
     lying
    0.87
     Lie
    0.82
     Lies
    0.78
    Lie
    0.74
    lie
    0.70
     lied
    0.66
    lies
    0.64
     liar
    0.64
    Activations Density 0.101%

    No Known Activations