© 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. 2-LLAMASCOPE-RES-32K
    11. 20331
    Prev
    Next
    INDEX
    Explanations

    phrases containing the word "off"

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With LLAMA3.1-8B @ 2-llamascope-res-32k
    Configuration
    fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L2R-8x
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    cerebras/SlimPajama-627B
    Features
    32,768
    Data Type
    bfloat16
    Hook Name
    blocks.2.hook_resid_post
    Hook Layer
    2
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    cerebras/SlimPajama-627B
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ever
    -0.15
     Piet
    -0.14
    zim
    -0.14
    ÏĦÏīν
    -0.14
     Riders
    -0.14
     Palestin
    -0.14
     crit
    -0.13
    illin
    -0.13
    Äħż
    -0.13
    agu
    -0.13
    POSITIVE LOGITS
    ischer
    0.15
    igne
    0.15
     Timing
    0.14
    rava
    0.14
    211
    0.14
    .Abstractions
    0.13
     trá»±c
    0.13
    pter
    0.13
    richt
    0.13
    imports
    0.13
    Activations Density 0.007%

    No Known Activations