© 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. 8214
    Prev
    Next
    INDEX
    Explanations

    technical and descriptive terms related to disc sports, such as discs, drivers, plastic material, and disc characteristics

    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
    698
    +0.14
    0.6%
    11
    +0.13
    0.6%
    391
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    698
    +0.14
    0.04
    5
    +0.13
    0.03
    506
    +0.11
    0.03
    Negative Logits
    \-
    -0.87
    Lua
    -0.84
    dl
    -0.81
    hex
    -0.76
    odi
    -0.76
    ril
    -0.75
    esc
    -0.75
     cf
    -0.74
    EE
    -0.73
    nel
    -0.73
    POSITIVE LOGITS
     MVP
    0.88
    Legend
    0.87
     Champion
    0.84
     Champions
    0.82
     Hero
    0.80
     Legend
    0.78
     superstar
    0.76
     STAR
    0.76
    Pros
    0.76
    Hero
    0.75
    Activations Density 0.194%

    No Known Activations