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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 20026
    Prev
    Next
    INDEX
    Explanations

    quotes or emphasized words

    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
    385
    +0.11
    0.5%
    314
    +0.11
    0.5%
    108
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    108
    +0.11
    0.02
    385
    +0.11
    0.02
    348
    +0.11
    0.02
    Negative Logits
     conj
    -0.68
     listed
    -0.67
     Vaugh
    -0.65
     scheduled
    -0.63
     paired
    -0.62
     Pavel
    -0.61
    azo
    -0.61
     statistic
    -0.61
     favor
    -0.61
     venture
    -0.60
    POSITIVE LOGITS
    Cause
    1.29
    Very
    1.07
    Absolutely
    1.07
    Interstitial
    1.02
    Extreme
    1.01
    cause
    1.00
    Really
    0.96
    Tw
    0.96
    They
    0.93
    Better
    0.93
    Activations Density 0.020%

    No Known Activations