© 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. 54773
    Prev
    Next
    INDEX
    Explanations

    conditional statements and references to occupation or settlement

    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
    mega
    -0.18
    ives
    -0.17
    ual
    -0.17
    ish
    -0.17
    ive
    -0.17
    eral
    -0.17
    ality
    -0.16
    sel
    -0.16
     Sel
    -0.16
    um
    -0.16
    POSITIVE LOGITS
    ying
    0.48
    ied
    0.40
    IED
    0.37
    ies
    0.31
    IES
    0.30
    iers
    0.30
    cation
    0.26
    ication
    0.25
    ier
    0.24
    ystate
    0.21
    Activations Density 0.007%

    No Known Activations