© 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. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 24576
    9. 8-RES_FS24576-JB
    11. 20104
    Prev
    Next
    INDEX
    Explanations

    verbs related to working or living, as well as adjectives related to maintaining, making, or being precise

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs24576-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    Data Type
    torch.float32
    Hook Point
    blocks.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    665
    +0.13
    0.6%
    70
    +0.13
    0.6%
    664
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    665
    +0.13
    0.07
    70
    +0.13
    0.07
    761
    +0.11
    0.05
    Negative Logits
    æľ
    -0.70
    Org
    -0.65
     wont
    -0.62
    èĪ
    -0.60
    }"
    -0.59
    orst
    -0.59
    Rust
    -0.58
    ouring
    -0.58
    ãĥ¼ãĥĨãĤ£
    -0.57
    ãĥĻ
    -0.57
    POSITIVE LOGITS
     freely
    1.29
     safely
    1.11
     comfortably
    1.04
     efficiently
    0.94
     indefinitely
    0.93
     uninterrupted
    0.91
     confidently
    0.90
     easily
    0.87
     lawfully
    0.87
     seamlessly
    0.86
    Activations Density 0.328%

    No Known Activations