© 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. 17-LLAMASCOPE-RES-32K
    11. 12809
    Prev
    Next
    INDEX
    Explanations

    terms and phrases associated with danger and risk

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With LLAMA3.1-8B @ 17-llamascope-res-32k
    Configuration
    fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L17R-8x
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    cerebras/SlimPajama-627B
    Features
    32,768
    Data Type
    bfloat16
    Hook Name
    blocks.17.hook_resid_post
    Hook Layer
    17
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    cerebras/SlimPajama-627B
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ãģ¡ãĤĩ
    -0.16
    лÑı
    -0.15
    Äįan
    -0.15
    ebek
    -0.14
    ijn
    -0.14
    IFI
    -0.14
    better
    -0.14
    éĢł
    -0.14
    å±
    -0.14
    pdb
    -0.14
    POSITIVE LOGITS
    -danger
    0.23
    ously
    0.22
     dangerous
    0.20
    éļª
    0.19
    stell
    0.18
    ous
    0.17
     danger
    0.17
     dangers
    0.17
     unsafe
    0.17
     Danger
    0.16
    Activations Density 0.027%

    No Known Activations