Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. OpenMOSS · Llama Scope: SAEs for Llama-3.1-8B
    3. Llama3.1-8B (Base)
    4. Residual Stream
    5. 26-LLAMASCOPE-RES-32K
    6. 27908
    Prev
    Next
    INDEX
    Explanations

    expressions of inconvenience and frustration

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With LLAMA3.1-8B @ 26-llamascope-res-32k
    Configuration
    fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L26R-8x
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    cerebras/SlimPajama-627B
    Features
    32,768
    Data Type
    bfloat16
    Hook Name
    blocks.26.hook_resid_post
    Hook Layer
    26
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    cerebras/SlimPajama-627B
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    igo
    -0.15
    à¤Ĥध
    -0.15
    Ñħод
    -0.14
     weakened
    -0.14
     overwhelmed
    -0.14
    oug
    -0.13
    rello
    -0.13
    juries
    -0.13
    idual
    -0.13
     empowering
    -0.13
    POSITIVE LOGITS
     disturbance
    0.40
     bother
    0.40
     interrupt
    0.40
     disrupt
    0.40
    intr
    0.39
     intr
    0.39
     disruptive
    0.38
     disturb
    0.38
     annoyance
    0.37
     disruption
    0.37
    Activations Density 0.396%

    No Known Activations