© 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. OpenMOSS · Llama Scope: SAEs for Llama-3.1-8B
    5. Llama3.1-8B (Base)
    7. Residual Stream
    9. 31-LLAMASCOPE-RES-32K
    11. 2339
    Prev
    Next
    INDEX
    Explanations

    elements related to classic video games and gameplay mechanics

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With LLAMA3.1-8B @ 31-llamascope-res-32k
    Configuration
    fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L31R-8x
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    cerebras/SlimPajama-627B
    Features
    32,768
    Data Type
    bfloat16
    Hook Name
    blocks.31.hook_resid_post
    Hook Layer
    31
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    cerebras/SlimPajama-627B
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     tumblr
    -0.22
     reddit
    -0.22
     Reddit
    -0.22
     subreddit
    -0.21
     Tumblr
    -0.20
     Pinterest
    -0.20
     instagram
    -0.19
     pinterest
    -0.19
     WhatsApp
    -0.19
     Airbnb
    -0.18
    POSITIVE LOGITS
     graphics
    0.18
     Graphics
    0.17
     FM
    0.17
     controls
    0.16
     VGA
    0.16
     CD
    0.16
     arcade
    0.16
     control
    0.16
     Cairo
    0.15
     levels
    0.15
    Activations Density 0.021%

    No Known Activations