© 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. 25-LLAMASCOPE-RES-32K
    11. 19196
    Prev
    Next
    INDEX
    Explanations

    references to places, particularly in Hawaii

    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
    aben
    -0.14
    има
    -0.14
    ignon
    -0.14
    irs
    -0.14
    ův
    -0.14
    /MPL
    -0.14
    orsk
    -0.14
     GANG
    -0.14
    edia
    -0.13
    -Ta
    -0.13
    POSITIVE LOGITS
     Punch
    0.27
     Wa
    0.23
     Pearl
    0.22
     Honolulu
    0.22
     Hick
    0.21
    Pear
    0.20
     Hale
    0.20
     Ala
    0.19
     wa
    0.19
     Diamond
    0.19
    Activations Density 0.006%

    No Known Activations