© 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 49152
    9. 8-RES_FS49152-JB
    11. 37335
    Prev
    Next
    INDEX
    Explanations

    punctuation and sentence structures

    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
    +0.20
    1.0%
    6
    +0.13
    0.6%
    679
    +0.13
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    6
    +0.20
    0.09
    679
    +0.13
    0.07
    640
    +0.13
    0.07
    Negative Logits
    urance
    -0.86
    plet
    -0.85
     crunch
    -0.83
    izer
    -0.79
     footing
    -0.77
     paycheck
    -0.76
    ierre
    -0.76
     mosqu
    -0.74
    omics
    -0.74
     damned
    -0.72
    POSITIVE LOGITS
     However
    1.57
     Additionally
    1.50
     Interestingly
    1.48
     Later
    1.44
     Also
    1.44
     Although
    1.41
     Alternatively
    1.40
     Ironically
    1.39
     Nevertheless
    1.39
     Furthermore
    1.38
    Activations Density 0.266%

    No Known Activations