© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Julius Han · Gated SAE for Llama3-8B-Instruct
    5. Llama3-8B-IT
    7. Residual Stream
    9. 25-RES-JH
    11. 23882
    Prev
    Next
    INDEX
    Explanations

    elements of deception and betrayal in narrative contexts

    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
    ÑĢана
    -0.15
    putc
    -0.15
    opyright
    -0.15
    Touches
    -0.14
    uga
    -0.14
     Lorenzo
    -0.14
    wdx
    -0.14
     Antar
    -0.13
     má»ĩnh
    -0.13
    ÄĽnÃŃ
    -0.13
    POSITIVE LOGITS
     trap
    0.89
     traps
    0.79
     Trap
    0.71
    trap
    0.69
    Trap
    0.66
     trapping
    0.60
     bait
    0.51
    _trap
    0.50
     trapped
    0.42
     ambush
    0.41
    Activations Density 0.025%

    No Known Activations