© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. OpenMOSS · Llama Scope: SAEs for Llama-3.1-8B
    3. Llama3.1-8B (Base)
    4. Residual Stream
    5. 31-LLAMASCOPE-RES-32K
    6. 17208
    Prev
    Next
    INDEX
    Explanations

    themes related to rescue and return to safety

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With LLAMA3.1-8B @ 31-llamascope-res-32k
    Configuration
    fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L31R-8x
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    cerebras/SlimPajama-627B
    Features
    32,768
    Data Type
    bfloat16
    Hook Name
    blocks.31.hook_resid_post
    Hook Layer
    31
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    cerebras/SlimPajama-627B
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     overposting
    -0.15
    Extern
    -0.14
    EXTERN
    -0.14
    .parallel
    -0.14
    жÑĥ
    -0.13
    raci
    -0.13
     universally
    -0.13
    inux
    -0.13
     extensively
    -0.13
    екÑĤ
    -0.12
    POSITIVE LOGITS
     alive
    0.49
     safely
    0.47
     safe
    0.42
    alive
    0.40
     Alive
    0.35
     intact
    0.33
    Alive
    0.32
    safe
    0.31
     safety
    0.31
    _safe
    0.30
    Activations Density 0.189%

    No Known Activations