© 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
    3. OpenMOSS · Llama Scope: SAEs for Llama-3.1-8B
    5. Llama3.1-8B (Base)
    7. Residual Stream
    9. 29-LLAMASCOPE-RES-32K
    11. 18624
    Prev
    Next
    INDEX
    Explanations

    negative attributes and behaviors associated with characters, particularly cruelty and abrasiveness

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With LLAMA3.1-8B @ 29-llamascope-res-32k
    Configuration
    fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L29R-8x
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    cerebras/SlimPajama-627B
    Features
    32,768
    Data Type
    bfloat16
    Hook Name
    blocks.29.hook_resid_post
    Hook Layer
    29
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    cerebras/SlimPajama-627B
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ovice
    -0.17
    eteria
    -0.17
    EMPL
    -0.16
     Bedford
    -0.15
    ekten
    -0.14
    laughter
    -0.14
    830
    -0.14
     khung
    -0.14
     Zwe
    -0.14
    owi
    -0.13
    POSITIVE LOGITS
     mean
    0.33
     abrasive
    0.33
     comb
    0.32
     ob
    0.30
     Mean
    0.29
    mean
    0.29
     rude
    0.28
     alo
    0.28
     confront
    0.27
     entitled
    0.27
    Activations Density 0.478%

    No Known Activations