© 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
    2. Julius Han · Gated SAE for Llama3-8B-Instruct
    3. Llama3-8B-IT
    4. Residual Stream
    5. 25-RES-JH
    6. 33536
    Prev
    Next
    INDEX
    Explanations

    instances of emotional reactions and descriptors related to experiences

    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
     stabilized
    -0.17
     marvel
    -0.16
     stabilize
    -0.16
    ipple
    -0.14
    otropic
    -0.14
    â
    -0.14
    lude
    -0.14
     Huffman
    -0.14
     stabilization
    -0.14
    ropa
    -0.14
    POSITIVE LOGITS
    crap
    0.18
     rubbish
    0.18
     emot
    0.17
    combe
    0.17
     envis
    0.17
     erm
    0.16
     whilst
    0.15
    Sorted
    0.15
    shit
    0.15
     specialist
    0.15
    Activations Density 0.106%

    No Known Activations