© 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. 7516
    Prev
    Next
    INDEX
    Explanations

    references to reports, radio shows, and subscription services

    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
    481
    +0.14
    0.6%
    39
    +0.11
    0.5%
    723
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    388
    +0.14
    0.05
    39
    +0.11
    0.05
    267
    +0.09
    0.05
    Negative Logits
     hurd
    -0.70
    )=
    -0.67
     nonexistent
    -0.66
     hump
    -0.66
     plant
    -0.64
     substituted
    -0.63
     strokes
    -0.61
     disqual
    -0.61
     draining
    -0.61
     obstruct
    -0.61
    POSITIVE LOGITS
     Podcast
    1.19
     Forums
    1.11
     Presents
    1.07
     Magazine
    1.05
     Wiki
    1.00
     Labs
    0.97
     Newsletter
    0.96
     Blog
    0.96
     Productions
    0.95
     Premium
    0.93
    Activations Density 0.357%

    No Known Activations