© 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 12288
    9. 8-RES_FS12288-JB
    11. 10827
    Prev
    Next
    INDEX
    Explanations

    phrases related to music and performance

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs12288-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    12,288
    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
    +0.19
    0.9%
    6
    +0.15
    0.7%
    679
    +0.13
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    6
    +0.19
    0.14
    640
    +0.15
    0.11
    679
    +0.13
    0.11
    Negative Logits
     legally
    -0.93
     authorized
    -0.92
     credential
    -0.89
     emergency
    -0.89
     legal
    -0.88
     renegoti
    -0.87
     prohibited
    -0.85
     lease
    -0.84
     voluntarily
    -0.83
     allocation
    -0.83
    POSITIVE LOGITS
     Featuring
    1.45
     Characters
    1.37
     Whereas
    1.33
     Though
    1.31
     Scenes
    1.31
     Thankfully
    1.27
     Regardless
    1.24
     Throughout
    1.24
     Along
    1.23
     Nonetheless
    1.22
    Activations Density 0.593%

    No Known Activations