© 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. 24032
    Prev
    Next
    INDEX
    Explanations

    specific locations and events related to military, familial, or significant social contexts

    oai_token-act-pair · gpt-4o-miniTriggered 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
    87
    +0.22
    1.1%
    633
    +0.09
    0.5%
    476
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    363
    +0.22
    0.05
    288
    +0.09
    0.09
    575
    +0.09
    0.07
    Negative Logits
    zsche
    -0.58
     voices
    -0.57
     relates
    -0.56
     mete
    -0.54
     Quote
    -0.53
     envy
    -0.53
     tsun
    -0.52
     summed
    -0.52
     fingerprints
    -0.52
     persists
    -0.52
    POSITIVE LOGITS
     premises
    0.89
     kios
    0.87
     station
    0.86
     checkpoint
    0.86
     headquarters
    0.86
     plantation
    0.83
     facility
    0.82
     encamp
    0.80
     shrine
    0.78
     plaza
    0.78
    Activations Density 0.914%

    No Known Activations