© 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. 23924
    Prev
    Next
    INDEX
    Explanations

    conjunctions and relative clauses indicating a defining relationship

    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
    679
    +0.14
    0.6%
    655
    +0.13
    0.6%
    393
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    655
    +0.14
    0.08
    109
    +0.13
    0.06
    551
    +0.11
    0.06
    Negative Logits
    ollo
    -0.70
    atre
    -0.61
    aq
    -0.60
     Basics
    -0.60
     Doctrine
    -0.60
    MX
    -0.60
    ball
    -0.59
    resa
    -0.59
    rou
    -0.59
     guessing
    -0.58
    POSITIVE LOGITS
     extends
    1.19
     justifies
    1.15
     accompanies
    1.14
     culmin
    1.13
     spans
    1.10
     exceeds
    1.10
     includes
    1.09
     outweigh
    1.07
     enhances
    1.07
     enables
    1.05
    Activations Density 0.194%

    No Known Activations