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

    Neuronpedia

    Jacobian LensNEW
    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. 6297
    Prev
    Next
    INDEX
    Explanations

    technical specifications and performance reviews related to computer hardware components

    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
    193
    +0.11
    0.5%
    620
    +0.10
    0.5%
    749
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    620
    +0.11
    0.02
    566
    +0.10
    0.05
    526
    +0.09
    -0.01
    Negative Logits
    lihood
    -0.76
     withdrew
    -0.72
     renegoti
    -0.71
     Fixed
    -0.71
     Procedures
    -0.69
    ļéĨĴ
    -0.68
     filing
    -0.67
    specified
    -0.67
     reverted
    -0.64
    Downloadha
    -0.64
    POSITIVE LOGITS
     unforgettable
    1.37
     rave
    1.12
     unbeat
    1.12
     classics
    1.11
     unparalleled
    1.05
     breathtaking
    1.04
     memorable
    1.03
     awe
    1.03
     thrilling
    1.02
     unmatched
    1.02
    Activations Density 0.819%

    No Known Activations