© 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 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. 14096
    Prev
    Next
    INDEX
    Explanations

    expressions of future potential and confidence in individuals’ abilities

    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
    ahren
    -0.15
    oud
    -0.15
    izard
    -0.15
    gba
    -0.15
    iti
    -0.14
    -alist
    -0.14
    ITCH
    -0.14
    ITU
    -0.14
    .ml
    -0.14
     Cle
    -0.14
    POSITIVE LOGITS
     assets
    0.24
     asset
    0.24
     Asset
    0.20
     Assets
    0.20
    assets
    0.20
     add
    0.18
     adds
    0.18
    Assets
    0.18
    ythe
    0.18
     excel
    0.17
    Activations Density 0.025%

    No Known Activations