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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsBlog/PodcastSlackPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlog/PodcastGitHubSlackTwitterContact
    1. Home
    2. Julius Han · Gated SAE for Llama3-8B-Instruct
    3. Llama3-8B-IT
    4. Residual Stream
    5. 25-RES-JH
    6. 23616
    Prev
    Next
    INDEX
    Explanations

    references to iconic villain characters in animated films

    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

    Negative Logits
    esty
    -0.16
    riz
    -0.15
    éĨ´
    -0.15
     irres
    -0.15
     çij
    -0.15
     embarrass
    -0.14
    Sac
    -0.14
    gii
    -0.14
     mirac
    -0.14
    isay
    -0.14
    POSITIVE LOGITS
     evil
    0.49
     Evil
    0.41
     villain
    0.39
    evil
    0.34
     Vill
    0.31
     villains
    0.31
     vill
    0.30
     sinister
    0.29
     malignant
    0.28
     male
    0.27
    Activations Density 0.086%

    No Known Activations