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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 41239
    Prev
    Next
    INDEX
    Explanations

    references to tactics, strategies, and the behavior of individuals in contexts of conflict or deception

    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
     impro
    -0.18
    rimp
    -0.15
     imp
    -0.15
     traj
    -0.14
    rip
    -0.13
    erre
    -0.13
     impr
    -0.13
     juxtap
    -0.13
    posium
    -0.13
    ãĤ¯
    -0.12
    POSITIVE LOGITS
     tactics
    0.35
     strategy
    0.30
     Tactics
    0.30
     tactic
    0.30
     techniques
    0.27
     strategies
    0.26
     technique
    0.26
     playbook
    0.26
     Strategy
    0.23
     successful
    0.23
    Activations Density 0.043%

    No Known Activations