© 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. 29536
    Prev
    Next
    INDEX
    Explanations

    phrases indicating refusal or avoidance of answering questions

    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
    etto
    -0.19
    ADDE
    -0.16
    stile
    -0.16
    oa
    -0.15
    feit
    -0.15
    _locals
    -0.15
    ì·¨
    -0.14
    éĿ©
    -0.14
    MESS
    -0.14
     Kore
    -0.14
    POSITIVE LOGITS
    won
    0.26
     won
    0.22
    Won
    0.21
     Won
    0.20
     wouldn
    0.18
     wont
    0.18
     not
    0.17
    ingly
    0.17
     tel
    0.16
    later
    0.16
    Activations Density 0.015%

    No Known Activations