© 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. 13181
    Prev
    Next
    INDEX
    Explanations

    technical terms related to hardware and software, like sensors, cameras, and software configurations

    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
    765
    +0.11
    0.5%
    593
    +0.11
    0.5%
    714
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    36
    +0.11
    0.05
    362
    +0.11
    0.05
    765
    +0.09
    0.04
    Negative Logits
    bernatorial
    -0.87
    worthiness
    -0.82
    iferation
    -0.70
    akings
    -0.69
     affair
    -0.68
    ependence
    -0.68
     justifying
    -0.68
    ocaust
    -0.68
    llor
    -0.67
    jobs
    -0.67
    POSITIVE LOGITS
     technique
    0.98
     techniques
    0.98
     pseudonym
    0.94
     combination
    0.92
     borrowed
    0.90
     methods
    0.90
     proprietary
    0.88
     methodology
    0.85
     terminology
    0.84
     method
    0.82
    Activations Density 0.233%

    No Known Activations