© 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 12288
    9. 8-RES_FS12288-JB
    11. 9517
    Prev
    Next
    INDEX
    Explanations

    references to security and digital information concerns

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs12288-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    12,288
    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
    +0.29
    1.4%
    393
    +0.20
    1.0%
    362
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    362
    +0.29
    0.10
    385
    +0.20
    0.10
    393
    +0.10
    0.09
    Negative Logits
    icia
    -0.82
    ibaba
    -0.76
    strate
    -0.74
    Alert
    -0.74
    ornings
    -0.72
    NB
    -0.71
    thia
    -0.70
    cellaneous
    -0.70
    ãĤ´ãĥ³
    -0.70
    bg
    -0.69
    POSITIVE LOGITS
     slightest
    1.22
     majority
    1.12
     entire
    1.04
     vast
    0.99
     greatest
    0.97
     Russians
    0.92
     absence
    0.91
     biggest
    0.88
     whole
    0.87
     latter
    0.87
    Activations Density 0.582%

    No Known Activations