© 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. 32928
    Prev
    Next
    INDEX
    Explanations

    references to power dynamics and shifts in authority

    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
    itta
    -0.15
    alar
    -0.15
    277
    -0.15
    ipa
    -0.15
    mailbox
    -0.14
    mot
    -0.14
    adia
    -0.14
    ÏĪε
    -0.14
    绩
    -0.14
    ismic
    -0.13
    POSITIVE LOGITS
     power
    0.94
    power
    0.76
     Power
    0.72
    -power
    0.69
    Power
    0.66
     POWER
    0.66
    _power
    0.61
    POWER
    0.58
    (power
    0.57
    .power
    0.56
    Activations Density 0.048%

    No Known Activations