© 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. 23997
    Prev
    Next
    INDEX
    Explanations

    references to rules or boundaries, particularly in legal 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
    362
    +0.16
    0.7%
    393
    +0.13
    0.6%
    501
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    362
    +0.16
    0.05
    393
    +0.13
    0.03
    635
    +0.12
    0.02
    Negative Logits
    behind
    -0.73
    MRI
    -0.72
     ahead
    -0.70
    PLA
    -0.68
    CLA
    -0.68
    DAY
    -0.67
     differently
    -0.66
    olkien
    -0.65
     anew
    -0.65
    PsyNetMessage
    -0.65
    POSITIVE LOGITS
     horizon
    1.18
     threshold
    1.13
     bounds
    1.08
     confines
    1.00
     boundaries
    0.96
     fray
    0.94
     plateau
    0.92
     usual
    0.91
     borders
    0.87
     surface
    0.85
    Activations Density 0.099%

    No Known Activations