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

    Neuronpedia

    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. 25-LLAMASCOPE-RES-32K
    11. 17420
    Prev
    Next
    INDEX
    Explanations

    technical syntax and structure in programming code

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

    Looking at the activations, this neuron fires strongly on programming syntax elements, particularly: - Assignment operators and equals signs (=) - Type specifications and declarations (CGSizeMake, CLLocationCoordinate2D) - Method parameters and function calls with colons

    oai_token-act-pair · claude-4-5-sonnetTriggered by @jyhe0408

    Objective-C style code patterns, especially CamelCase API/type identifiers, property/method declarations with colons and parentheses, and assignment/comparison operations.

    oai_token-act-pair · gpt-5Triggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With LLAMA3.1-8B @ 25-llamascope-res-32k
    Configuration
    fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L25R-8x
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    cerebras/SlimPajama-627B
    Features
    32,768
    Data Type
    bfloat16
    Hook Name
    blocks.25.hook_resid_post
    Hook Layer
    25
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    cerebras/SlimPajama-627B
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ucs
    -0.17
    _UC
    -0.15
     delegate
    -0.15
    kest
    -0.15
    ucid
    -0.15
    icans
    -0.14
    icut
    -0.14
    ishes
    -0.14
     outfit
    -0.14
     coolant
    -0.14
    POSITIVE LOGITS
     CG
    0.28
    CG
    0.28
    cg
    0.26
    .CG
    0.25
    (CG
    0.23
     cg
    0.22
    _cg
    0.21
    :CGRect
    0.19
     Unsafe
    0.18
     CF
    0.17
    Activations Density 0.007%

    No Known Activations