Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsBlog/PodcastSlackPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlog/PodcastGitHubSlackTwitterContact
    1. Home
    2. Julius Han · Gated SAE for Llama3-8B-Instruct
    3. Llama3-8B-IT
    4. Residual Stream
    5. 25-RES-JH
    6. 23605
    Prev
    Next
    INDEX
    Explanations

    descriptions of physical measurements related to elevation and location

    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
     góc
    -0.15
     Hod
    -0.14
     walls
    -0.14
     Gro
    -0.14
    еÑĢж
    -0.14
    unge
    -0.13
     æĴ
    -0.13
     |
    -0.13
     apart
    -0.13
     >
    -0.13
    POSITIVE LOGITS
     level
    0.39
     ground
    0.36
    ground
    0.30
     sea
    0.29
     grade
    0.29
    level
    0.27
     nivel
    0.27
     Level
    0.26
    Level
    0.25
    -ground
    0.25
    Activations Density 0.028%

    No Known Activations