© 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 EvalsExportsinterp-engineNEWAPI 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. 22215
    Prev
    Next
    INDEX
    Explanations

    references to cleaning processes or methods

    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
     liqu
    -0.14
     hyp
    -0.14
     unp
    -0.14
    dirty
    -0.14
     mur
    -0.14
    atrix
    -0.14
     washer
    -0.13
     dissolved
    -0.13
    chs
    -0.13
    εÏģγ
    -0.13
    POSITIVE LOGITS
     drying
    0.54
     dried
    0.41
     dryer
    0.38
     dry
    0.37
     Dry
    0.36
    Dry
    0.35
    dry
    0.34
     خشک
    0.25
    ä¹¾
    0.25
    å¹²
    0.24
    Activations Density 0.027%

    No Known Activations