© 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. 41004
    Prev
    Next
    INDEX
    Explanations

    instances of the verb "was" and related forms to highlight past events and conditions

    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
    NOW
    -0.20
     ><?
    -0.16
    elter
    -0.15
    now
    -0.14
    -now
    -0.14
    anka
    -0.14
    akah
    -0.14
    (now
    -0.14
    bots
    -0.14
    hari
    -0.14
    POSITIVE LOGITS
     pronounced
    0.22
     crem
    0.21
     last
    0.20
    pron
    0.20
     Last
    0.18
     Pron
    0.18
     buried
    0.17
     survived
    0.17
     remembered
    0.17
     killed
    0.17
    Activations Density 0.031%

    No Known Activations