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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 8737
    Prev
    Next
    INDEX
    Explanations

    references to provocative or controversial topics, including elements of horror or societal critique

    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
    -Semit
    -0.15
    اÙĨÙĩ
    -0.14
     gần
    -0.14
     lety
    -0.14
    agli
    -0.14
    ãĨ
    -0.14
     Reyes
    -0.13
    istrovstvÃŃ
    -0.13
    /licenses
    -0.13
    ëį°ìĿ´íĬ¸
    -0.13
    POSITIVE LOGITS
     tomorrow
    0.20
    and
    0.17
    znam
    0.17
    while
    0.16
    aming
    0.15
     backwards
    0.15
     probably
    0.14
     type
    0.14
     preferably
    0.14
    ìĮ
    0.14
    Activations Density 0.073%

    No Known Activations