© 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. 30055
    Prev
    Next
    INDEX
    Explanations

    instances of the word "allow" and its variations, particularly in contexts discussing permissions and restrictions

    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
    /from
    -0.17
    ugar
    -0.15
    .dp
    -0.14
    omu
    -0.14
    borg
    -0.14
    installer
    -0.14
    .Throw
    -0.14
    urette
    -0.14
    .DefaultCellStyle
    -0.14
    iro
    -0.13
    POSITIVE LOGITS
     themselves
    0.21
     ourselves
    0.20
     continued
    0.19
    fullscreen
    0.18
     continue
    0.17
     himself
    0.17
     room
    0.17
     slip
    0.17
     continuation
    0.17
     latitude
    0.17
    Activations Density 0.057%

    No Known Activations