© 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 49152
    9. 8-RES_FS49152-JB
    11. 26757
    Prev
    Next
    INDEX
    Explanations

    punctuation marks, particularly periods, exclamation marks, and question marks

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs49152-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    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
    640
    +0.18
    0.9%
    6
    +0.14
    0.7%
    679
    +0.14
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    6
    +0.18
    0.15
    679
    +0.14
    0.12
    640
    +0.14
    0.10
    Negative Logits
     emanc
    -0.83
     glim
    -0.81
     sequential
    -0.71
     tractor
    -0.71
     succession
    -0.70
     defe
    -0.70
     enrol
    -0.70
     cradle
    -0.69
     involuntary
    -0.69
     rall
    -0.69
    POSITIVE LOGITS
     Hence
    1.31
     Therefore
    1.29
     Thus
    1.24
     Moreover
    1.21
     Furthermore
    1.19
     Regardless
    1.16
     However
    1.16
     Accordingly
    1.16
     Especially
    1.15
     Additionally
    1.14
    Activations Density 0.509%

    No Known Activations