© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. OpenMOSS · Llama Scope: SAEs for Llama-3.1-8B
    5. Llama3.1-8B (Base)
    7. Residual Stream
    9. 16-LLAMASCOPE-RES-32K
    11. 4276
    Prev
    Next
    INDEX
    Explanations

    access modifiers and method declarations in Java-like programming languages

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With LLAMA3.1-8B @ 16-llamascope-res-32k
    Configuration
    fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L16R-8x
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    cerebras/SlimPajama-627B
    Features
    32,768
    Data Type
    bfloat16
    Hook Name
    blocks.16.hook_resid_post
    Hook Layer
    16
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    cerebras/SlimPajama-627B
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    oser
    -0.19
    STANCE
    -0.16
    roperties
    -0.14
    odÃŃ
    -0.14
    ãĥ³ãĤº
    -0.14
    alsex
    -0.14
     onBackPressed
    -0.14
    irma
    -0.14
    idak
    -0.14
    ston
    -0.13
    POSITIVE LOGITS
    445
    0.17
    گاÙĩ
    0.16
    arget
    0.16
    ä¹İ
    0.15
    ors
    0.15
    347
    0.15
    699
    0.15
    346
    0.15
    olist
    0.14
     Boone
    0.14
    Activations Density 0.006%

    No Known Activations