© 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. 41657
    Prev
    Next
    INDEX
    Explanations

    references to the Supreme Court and its members, particularly focusing on appointments and their implications

    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
    hire
    -0.18
    umb
    -0.17
     magn
    -0.15
    rado
    -0.15
    rin
    -0.15
    JKLMNOP
    -0.14
    dek
    -0.14
    à¹Ģà¸Ń
    -0.14
     Congressman
    -0.14
    Justice
    -0.14
    POSITIVE LOGITS
     bench
    0.45
    bench
    0.42
     Bench
    0.40
     benches
    0.27
     Supreme
    0.21
    benchmark
    0.21
     benchmark
    0.21
    Benchmark
    0.20
     Court
    0.18
     Benchmark
    0.17
    Activations Density 0.030%

    No Known Activations