© 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. 21811
    Prev
    Next
    INDEX
    Explanations

    phrases indicating coincidence or unexpected occurrence

    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
     Giang
    -0.17
    wd
    -0.16
    urtles
    -0.15
    μι
    -0.15
    dit
    -0.14
    acket
    -0.14
     JADX
    -0.14
    ê°Ħ
    -0.13
    ethyst
    -0.13
    pg
    -0.13
    POSITIVE LOGITS
     happen
    0.93
     happens
    0.91
     happened
    0.77
     Happ
    0.71
     happening
    0.68
     happ
    0.58
     aconte
    0.51
     HAPP
    0.44
     coincidence
    0.38
     coinc
    0.36
    Activations Density 0.031%

    No Known Activations