© 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 24576
    9. 8-RES_FS24576-JB
    11. 6852
    Prev
    Next
    INDEX
    Explanations

    ethical principles and instructions for conduct

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs24576-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    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
    87
    +0.29
    1.4%
    459
    +0.09
    0.5%
    285
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    87
    +0.29
    -0.02
    285
    +0.09
    0.05
    756
    +0.09
    -0.02
    Negative Logits
     leaked
    -0.64
    UPDATE
    -0.62
     Sample
    -0.61
    MpServer
    -0.61
     crashed
    -0.60
    soon
    -0.60
    reported
    -0.60
     announced
    -0.59
    arthed
    -0.59
     Reported
    -0.58
    POSITIVE LOGITS
     humility
    1.04
     loyalty
    1.01
     persever
    1.01
     teamwork
    0.99
     obedience
    0.96
     honesty
    0.96
    cellence
    0.95
     vigilance
    0.95
     dignity
    0.94
     strive
    0.93
    Activations Density 0.749%

    No Known Activations