© 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. 41311
    Prev
    Next
    INDEX
    Explanations

    expressions emphasizing the pronouns "it" and "that"

    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
    æº
    -0.17
    rq
    -0.16
    ห
    -0.16
    ahl
    -0.16
    303
    -0.15
    prev
    -0.13
    avra
    -0.13
    quito
    -0.12
    ups
    -0.12
    ips
    -0.12
    POSITIVE LOGITS
     is
    0.28
     was
    0.21
     ÑıвлÑıеÑĤÑģÑı
    0.20
     æĺ¯
    0.19
    æĺ¯
    0.18
     truly
    0.17
     Is
    0.17
    _is
    0.16
    ardır
    0.16
     are
    0.16
    Activations Density 0.080%

    No Known Activations