© 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. 41729
    Prev
    Next
    INDEX
    Explanations

    references to politically charged psychological terms and disorders

    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
    Stub
    -0.15
    ignet
    -0.14
     ApplicationException
    -0.14
    orca
    -0.14
    ugin
    -0.13
     ListViewItem
    -0.13
    anda
    -0.13
     Stafford
    -0.13
    .utc
    -0.13
    491
    -0.13
    POSITIVE LOGITS
     syndrome
    0.86
     Syndrome
    0.79
     synd
    0.67
     Synd
    0.63
    sy
    0.55
    Sy
    0.54
    _sy
    0.50
     SY
    0.46
     Sy
    0.44
     sy
    0.42
    Activations Density 0.014%

    No Known Activations