© 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. 18587
    Prev
    Next
    INDEX
    Explanations

    scientific and technical jargon related to outer space and extraterrestrial environments

    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
    724
    +0.12
    0.5%
    729
    +0.10
    0.4%
    689
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    311
    +0.12
    0.02
    724
    +0.10
    0.05
    242
    +0.10
    0.04
    Negative Logits
    ngth
    -0.96
    lished
    -0.72
    assemb
    -0.72
    ividual
    -0.71
    çīĪ
    -0.70
    rive
    -0.70
    ideshow
    -0.69
    iries
    -0.68
    orously
    -0.67
    ravings
    -0.66
    POSITIVE LOGITS
     heresy
    0.82
     blasphemy
    0.81
     understatement
    0.81
     territory
    0.80
     scenario
    0.78
     explan
    0.78
     semantics
    0.76
     shenan
    0.76
     coincidence
    0.75
     shorthand
    0.74
    Activations Density 0.273%

    No Known Activations