© 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. 41727
    Prev
    Next
    INDEX
    Explanations

    references to a specific product or brand, particularly "Bubblers."

    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
    232
    +0.12
    0.6%
    504
    +0.12
    0.5%
    603
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    232
    +0.12
    0.01
    326
    +0.12
    0.01
    603
    +0.11
    0.01
    Negative Logits
    IVES
    -0.88
    WARE
    -0.78
    IVE
    -0.77
    ALLY
    -0.72
    IFE
    -0.70
    sson
    -0.70
    CTV
    -0.70
    IGHTS
    -0.69
    CVE
    -0.69
    BOOK
    -0.68
    POSITIVE LOGITS
    bles
    1.34
    bub
    1.03
    bling
    0.94
    bly
    0.94
    bled
    0.91
     champagne
    0.86
    earance
    0.82
    blers
    0.81
    iltration
    0.81
    alore
    0.80
    Activations Density 0.005%

    No Known Activations