© 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. 3-LLAMASCOPE-RES-32K
    11. 3110
    Prev
    Next
    INDEX
    Explanations

    actions related to evasion or avoidance

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With LLAMA3.1-8B @ 3-llamascope-res-32k
    Configuration
    fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L3R-8x
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    cerebras/SlimPajama-627B
    Features
    32,768
    Data Type
    bfloat16
    Hook Name
    blocks.3.hook_resid_post
    Hook Layer
    3
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    cerebras/SlimPajama-627B
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    енд
    -0.16
    elib
    -0.15
     Ø´Ùĩ
    -0.15
    #ad
    -0.15
    .Extension
    -0.14
    nict
    -0.14
    urovision
    -0.14
    keiten
    -0.14
    serter
    -0.14
    меÑĩ
    -0.14
    POSITIVE LOGITS
     avoidance
    0.17
     avoided
    0.16
    avoid
    0.16
     Avoid
    0.15
    abox
    0.15
     recess
    0.15
    JD
    0.15
    jd
    0.15
    955
    0.15
     avoids
    0.15
    Activations Density 0.119%

    No Known Activations