© 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. 27-LLAMASCOPE-RES-32K
    11. 3871
    Prev
    Next
    INDEX
    Explanations

    references to flaws, errors, and character weaknesses in various contexts

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With LLAMA3.1-8B @ 27-llamascope-res-32k
    Configuration
    fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L27R-8x
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    cerebras/SlimPajama-627B
    Features
    32,768
    Data Type
    bfloat16
    Hook Name
    blocks.27.hook_resid_post
    Hook Layer
    27
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    cerebras/SlimPajama-627B
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     rarity
    -0.15
    urch
    -0.13
    ç½
    -0.13
    dea
    -0.13
     Antar
    -0.13
     insult
    -0.12
    iker
    -0.12
     interruption
    -0.12
     Impossible
    -0.12
     unauthorized
    -0.12
    POSITIVE LOGITS
     faults
    0.46
     flaws
    0.45
     flaw
    0.43
     fault
    0.41
     weaknesses
    0.40
     Fault
    0.39
     defects
    0.39
    fault
    0.36
     holes
    0.35
     deficiencies
    0.34
    Activations Density 0.373%

    No Known Activations