© 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. 25-LLAMASCOPE-RES-32K
    11. 14096
    Prev
    Next
    INDEX
    Explanations

    references to light conditions and visual perceptions

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With LLAMA3.1-8B @ 25-llamascope-res-32k
    Configuration
    fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L25R-8x
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    cerebras/SlimPajama-627B
    Features
    32,768
    Data Type
    bfloat16
    Hook Name
    blocks.25.hook_resid_post
    Hook Layer
    25
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    cerebras/SlimPajama-627B
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Dark
    -0.21
     dark
    -0.19
     DARK
    -0.19
    Dark
    -0.17
     Black
    -0.16
    é»Ĵ
    -0.16
     darker
    -0.16
     sounding
    -0.16
     blacklist
    -0.16
    _dark
    -0.16
    POSITIVE LOGITS
     diff
    0.23
     wash
    0.22
     reflected
    0.22
     filtered
    0.20
    -filter
    0.20
    wash
    0.20
     filtering
    0.19
     Wash
    0.19
     shaft
    0.19
    ambient
    0.19
    Activations Density 0.072%

    No Known Activations