© 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. 54385
    Prev
    Next
    INDEX
    Explanations

    references to the concept of loss

    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
    aber
    -0.15
    rollo
    -0.15
    alamat
    -0.14
    bara
    -0.14
    abay
    -0.13
     Rail
    -0.13
    izable
    -0.13
    lant
    -0.13
    endants
    -0.13
     rail
    -0.13
    POSITIVE LOGITS
     of
    0.27
    -of
    0.22
    y
    0.20
    (es
    0.19
    _of
    0.17
    ses
    0.17
     Of
    0.17
    gesch
    0.16
     của
    0.15
    ãģĹãĤĥ
    0.15
    Activations Density 0.009%

    No Known Activations