© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 40597
    Prev
    Next
    INDEX
    Explanations

    high-frequency function words and punctuation marks

    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
    eron
    -0.19
    iam
    -0.15
    eded
    -0.14
    ¸
    -0.14
    eln
    -0.13
     evet
    -0.13
    verting
    -0.13
    (er
    -0.13
    ãģĬãĤĬ
    -0.13
    _##
    -0.13
    POSITIVE LOGITS
    alsex
    0.14
    rick
    0.14
    acier
    0.14
    алеж
    0.14
    acea
    0.13
    Ñĥж
    0.13
    reeze
    0.13
    alette
    0.13
    zilla
    0.13
    ptions
    0.13
    Activations Density 0.000%

    No Known Activations