© 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. 20-LLAMASCOPE-RES-32K
    11. 23457
    Prev
    Next
    INDEX
    Explanations

    references to user interactions and personal information regarding privacy and cookies

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With LLAMA3.1-8B @ 20-llamascope-res-32k
    Configuration
    fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L20R-8x
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    cerebras/SlimPajama-627B
    Features
    32,768
    Data Type
    bfloat16
    Hook Name
    blocks.20.hook_resid_post
    Hook Layer
    20
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    cerebras/SlimPajama-627B
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ãĥ³ãĤ¹
    -0.17
    .compiler
    -0.16
    anders
    -0.15
    paces
    -0.15
    FETCH
    -0.14
    ZA
    -0.14
    dots
    -0.14
    gren
    -0.14
    Bindable
    -0.14
    coma
    -0.14
    POSITIVE LOGITS
     browsing
    0.22
     interaction
    0.21
    interaction
    0.20
     browse
    0.19
     interactions
    0.19
     visiting
    0.18
     interact
    0.18
    orsch
    0.17
    input
    0.17
    -input
    0.17
    Activations Density 0.053%

    No Known Activations