© 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. 18-LLAMASCOPE-RES-32K
    11. 10959
    Prev
    Next
    INDEX
    Explanations

    actions related to evading danger or seeking safety

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With LLAMA3.1-8B @ 18-llamascope-res-32k
    Configuration
    fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L18R-8x
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    cerebras/SlimPajama-627B
    Features
    32,768
    Data Type
    bfloat16
    Hook Name
    blocks.18.hook_resid_post
    Hook Layer
    18
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    cerebras/SlimPajama-627B
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Trab
    -0.15
     необ
    -0.14
    ToWorld
    -0.14
    ÐĴС
    -0.14
    964
    -0.14
     rehab
    -0.14
    963
    -0.14
    ulfilled
    -0.13
    fab
    -0.13
     Rehab
    -0.13
    POSITIVE LOGITS
     escape
    0.48
     escaping
    0.40
     flee
    0.39
     Escape
    0.38
     fleeing
    0.36
     escapes
    0.36
     escaped
    0.35
    escape
    0.35
    éĢĥ
    0.34
     fled
    0.33
    Activations Density 0.255%

    No Known Activations