© 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
    2. Julius Han · Gated SAE for Llama3-8B-Instruct
    3. Llama3-8B-IT
    4. Residual Stream
    5. 25-RES-JH
    6. 29486
    Prev
    Next
    INDEX
    Explanations

    the letter "H" in various contexts

    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
    ero
    -0.18
    ضÙĦ
    -0.17
    its
    -0.16
    ook
    -0.16
    ahun
    -0.15
    andler
    -0.15
    ier
    -0.15
    ieri
    -0.15
    gnore
    -0.15
    ASH
    -0.15
    POSITIVE LOGITS
    iran
    0.17
    ikal
    0.16
    oft
    0.16
    oor
    0.16
    uum
    0.16
     pyl
    0.16
    esper
    0.15
    iola
    0.15
     Accom
    0.15
    asso
    0.15
    Activations Density 0.032%

    No Known Activations