© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI 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. 20607
    Prev
    Next
    INDEX
    Explanations

    scientific research publications related to environmental issues and cognitive 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_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
    313
    +0.12
    0.5%
    435
    +0.09
    0.4%
    717
    +0.08
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    313
    +0.12
    0.02
    631
    +0.09
    0.02
    424
    +0.08
    0.01
    Negative Logits
     brim
    -0.72
    atown
    -0.70
    netflix
    -0.70
    irl
    -0.69
    uca
    -0.69
    agine
    -0.66
     herds
    -0.65
     tears
    -0.64
     reckoning
    -0.64
    andi
    -0.63
    POSITIVE LOGITS
     Conference
    0.93
     Proceedings
    0.89
     Subcommittee
    0.79
     Decision
    0.71
     Transactions
    0.70
     congress
    0.68
     Appropriations
    0.67
    Reviewer
    0.67
     Experimental
    0.66
     sym
    0.66
    Activations Density 0.043%

    No Known Activations