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

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. OpenMOSS · Llama Scope: SAEs for Llama-3.1-8B
    5. Llama3.1-8B (Base)
    7. Residual Stream
    9. 6-LLAMASCOPE-RES-32K
    11. 19386
    Prev
    Next
    INDEX
    Explanations

    references to wrestling teams and personalities

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With LLAMA3.1-8B @ 6-llamascope-res-32k
    Configuration
    fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L6R-8x
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    cerebras/SlimPajama-627B
    Features
    32,768
    Data Type
    bfloat16
    Hook Name
    blocks.6.hook_resid_post
    Hook Layer
    6
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    cerebras/SlimPajama-627B
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ãĥĸãĥ«
    -0.17
    izon
    -0.16
     Blade
    -0.15
    .uc
    -0.14
     lij
    -0.14
    ãĤıãģĽ
    -0.14
     Ridge
    -0.14
     ÑģÑĤ
    -0.14
     Them
    -0.14
     Ä°z
    -0.14
    POSITIVE LOGITS
    adal
    0.16
     consisting
    0.15
    adas
    0.15
    .kwargs
    0.14
    347
    0.14
    auty
    0.14
    aley
    0.14
     tandem
    0.14
    åħ
    0.14
    hash
    0.14
    Activations Density 0.226%

    No Known Activations