© 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. 26765
    Prev
    Next
    INDEX
    Explanations

    instances of the verb "be" in various forms and contexts

    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
    314
    +0.12
    0.6%
    167
    +0.10
    0.4%
    706
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    102
    +0.12
    0.03
    55
    +0.10
    0.03
    706
    +0.09
    0.03
    Negative Logits
     Hert
    -0.63
     note
    -0.62
    fortunately
    -0.58
     refuse
    -0.58
     attempt
    -0.57
     Norn
    -0.57
    plex
    -0.57
     Kahn
    -0.56
     owe
    -0.56
     clerk
    -0.56
    POSITIVE LOGITS
     fun
    0.76
    acea
    0.75
     nonpartisan
    0.74
    fits
    0.73
     disruptive
    0.72
    eal
    0.72
    ocument
    0.69
    leeve
    0.69
     flexible
    0.67
     therapeutic
    0.67
    Activations Density 0.070%

    No Known Activations