Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. OpenMOSS · Llama Scope: SAEs for Llama-3.1-8B
    3. Llama3.1-8B (Base)
    4. Residual Stream
    5. 18-LLAMASCOPE-RES-32K
    6. 24632
    Prev
    Next
    INDEX
    Explanations

    instances of teaching and educational roles

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With LLAMA3.1-8B @ 18-llamascope-res-32k
    Configuration
    fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L18R-8x
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    cerebras/SlimPajama-627B
    Features
    32,768
    Data Type
    bfloat16
    Hook Name
    blocks.18.hook_resid_post
    Hook Layer
    18
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    cerebras/SlimPajama-627B
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    óm
    -0.15
    é½IJ
    -0.15
    ihan
    -0.15
    sci
    -0.15
    tel
    -0.15
    stantiate
    -0.14
    à¤Ł
    -0.14
    /msg
    -0.14
    uts
    -0.14
    ambi
    -0.14
    POSITIVE LOGITS
     classes
    0.19
     teach
    0.18
     teaching
    0.18
     teaches
    0.16
     classe
    0.15
    ¼
    0.15
     at
    0.14
     Manafort
    0.14
     Classes
    0.14
     clases
    0.14
    Activations Density 0.073%

    No Known Activations