© 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 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. 14469
    Prev
    Next
    INDEX
    Explanations

    words related to suffering or difficulties in various contexts

    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
    s
    -0.18
    sut
    -0.18
    lug
    -0.18
    sip
    -0.17
    tÃŃ
    -0.17
    leys
    -0.16
    ÑĢÑıдÑĥ
    -0.16
    echn
    -0.15
    elem
    -0.15
    sie
    -0.15
    POSITIVE LOGITS
    theless
    0.42
    ment
    0.34
    ity
    0.33
    gether
    0.33
    ation
    0.32
    ous
    0.32
    ian
    0.31
    ellaneous
    0.30
    istrator
    0.30
    ive
    0.30
    Activations Density 0.007%

    No Known Activations