© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 2647
    Prev
    Next
    INDEX
    Explanations

    patterns related to whitespace and string manipulation in code

    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
    ben
    -0.17
    usc
    -0.14
    Å¡Å¥
    -0.14
     èŤ
    -0.14
    -fashion
    -0.14
    regon
    -0.13
    olith
    -0.13
    636
    -0.13
    cept
    -0.13
    overs
    -0.13
    POSITIVE LOGITS
    unkt
    0.15
    aken
    0.15
    YY
    0.14
    nÃŃ
    0.14
     sag
    0.14
    Ranges
    0.13
    cobra
    0.13
    icontrol
    0.13
    .$$
    0.13
     Targets
    0.13
    Activations Density 0.040%

    No Known Activations