© 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. 22218
    Prev
    Next
    INDEX
    Explanations

    expressions of saying goodbye or farewell

    oai_token-act-pair · gpt-4o-miniTriggered 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
    726
    +0.11
    0.5%
    319
    +0.11
    0.5%
    759
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    242
    +0.11
    0.01
    759
    +0.11
    0.01
    144
    +0.10
    0.01
    Negative Logits
    rikes
    -0.94
    yrinth
    -0.80
    orum
    -0.77
    ris
    -0.74
    urg
    -0.73
    forces
    -0.72
    geries
    -0.71
    arist
    -0.71
    ickr
    -0.70
    é¾
    -0.70
    POSITIVE LOGITS
     goodbye
    1.21
    bye
    1.16
     Bye
    1.00
     bye
    0.92
     farewell
    0.92
     greet
    0.81
     Goodbye
    0.81
     kiss
    0.80
     hello
    0.79
     handshake
    0.77
    Activations Density 0.023%

    No Known Activations