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

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI 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. 22395
    Prev
    Next
    INDEX
    Explanations

    terms related to secrecy and classified information

    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
    že
    -0.14
    766
    -0.14
    åłĤ
    -0.13
    essional
    -0.13
    orical
    -0.13
    icine
    -0.13
    ems
    -0.13
    .promise
    -0.12
    esso
    -0.12
    unei
    -0.12
    POSITIVE LOGITS
     classified
    0.65
     Classified
    0.57
     secret
    0.55
    classified
    0.51
     top
    0.49
     Secret
    0.46
     SECRET
    0.45
     Top
    0.44
     classification
    0.43
     sensitive
    0.42
    Activations Density 0.058%

    No Known Activations