© 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. 8559
    Prev
    Next
    INDEX
    Explanations

    references to espionage and deception in relation to Israeli dynamics

    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
    insky
    -0.15
     chilling
    -0.14
    asma
    -0.14
    avig
    -0.14
    ÏĢά
    -0.14
    avage
    -0.14
    atÃŃm
    -0.14
     dodge
    -0.13
    517
    -0.13
    izzo
    -0.13
    POSITIVE LOGITS
     attack
    0.21
    fight
    0.20
     attacked
    0.19
    çĤ¸
    0.18
     bombing
    0.18
     fighting
    0.18
     peace
    0.18
     annex
    0.18
     liquid
    0.17
     fight
    0.17
    Activations Density 0.113%

    No Known Activations