© 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. 15-LLAMASCOPE-RES-32K
    11. 14298
    Prev
    Next
    INDEX
    Explanations

    phrases related to regret or apology

    oai_token-act-pair · gpt-4o-miniTriggered by @bot

    Looking at the activations, this neuron activates weakly (around 0.4-0.6) on common function words like "I'm", "sorry", "the", "and", "who", "was", "others", "grounds", and "war" that appear

    oai_token-act-pair · claude-4-5-sonnetTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With LLAMA3.1-8B @ 15-llamascope-res-32k
    Configuration
    fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L15R-8x
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    cerebras/SlimPajama-627B
    Features
    32,768
    Data Type
    bfloat16
    Hook Name
    blocks.15.hook_resid_post
    Hook Layer
    15
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    cerebras/SlimPajama-627B
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    inger
    -0.17
    cents
    -0.15
     ÄIJÃłi
    -0.14
     çİ
    -0.14
    asper
    -0.14
     ê³Ħ
    -0.14
    ìĽĶë¶ĢíĦ°
    -0.14
    μβ
    -0.14
    INGER
    -0.14
    olding
    -0.14
    POSITIVE LOGITS
    ILLE
    0.17
    éĺ
    0.16
    iets
    0.15
    èµ
    0.15
    èĮĤ
    0.15
    akes
    0.15
    bes
    0.15
     qu
    0.14
    ovice
    0.14
    리ì§Ģ
    0.14
    Activations Density 0.003%

    No Known Activations