© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. OpenMOSS · Llama Scope: SAEs for Llama-3.1-8B
    5. Llama3.1-8B (Base)
    7. Residual Stream
    9. 5-LLAMASCOPE-RES-32K
    11. 3841
    Prev
    Next
    INDEX
    Explanations

    expressions related to holding, grasping, or physically manipulating objects

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With LLAMA3.1-8B @ 5-llamascope-res-32k
    Configuration
    fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L5R-8x
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    cerebras/SlimPajama-627B
    Features
    32,768
    Data Type
    bfloat16
    Hook Name
    blocks.5.hook_resid_post
    Hook Layer
    5
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    cerebras/SlimPajama-627B
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    heten
    -0.17
    elow
    -0.16
    veau
    -0.16
    edList
    -0.15
    hq
    -0.14
    umpt
    -0.14
    PRECATED
    -0.14
    oted
    -0.14
     Couple
    -0.14
     womb
    -0.14
    POSITIVE LOGITS
     accountable
    0.20
    ä½ı
    0.19
     hold
    0.18
     Hold
    0.18
    Hold
    0.17
     ä½ı
    0.17
    rador
    0.17
    hold
    0.16
     HOLD
    0.16
     promise
    0.16
    Activations Density 0.065%

    No Known Activations