© 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. 20-LLAMASCOPE-RES-32K
    11. 13447
    Prev
    Next
    INDEX
    Explanations

    attributes related to physical appearance, particularly attractiveness

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With LLAMA3.1-8B @ 20-llamascope-res-32k
    Configuration
    fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L20R-8x
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    cerebras/SlimPajama-627B
    Features
    32,768
    Data Type
    bfloat16
    Hook Name
    blocks.20.hook_resid_post
    Hook Layer
    20
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    cerebras/SlimPajama-627B
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ylon
    -0.15
     Abstract
    -0.15
    aldo
    -0.15
    awi
    -0.14
     Activity
    -0.14
    inho
    -0.14
    itra
    -0.14
    mbH
    -0.14
     Actor
    -0.13
    atorium
    -0.13
    POSITIVE LOGITS
     sed
    0.23
     attractive
    0.23
     beauty
    0.21
     attractiveness
    0.20
     beautiful
    0.20
     hottest
    0.19
     gorgeous
    0.19
     beaut
    0.18
    ç¾İ
    0.18
     sexy
    0.18
    Activations Density 0.237%

    No Known Activations