© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW 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. 2986
    Prev
    Next
    INDEX
    Explanations

    references to academic and scientific research, and medical procedures or treatments

    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
    156
    +0.13
    0.6%
    632
    +0.10
    0.4%
    44
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    156
    +0.13
    0.04
    660
    +0.10
    0.05
    567
    +0.09
    0.04
    Negative Logits
    ãĥķãĤ©
    -0.81
    TPS
    -0.79
    veyard
    -0.78
    ionage
    -0.74
    Assembly
    -0.72
    rss
    -0.70
    Commission
    -0.70
    jl
    -0.69
    UTERS
    -0.68
    Chain
    -0.67
    POSITIVE LOGITS
     subconscious
    0.93
     unconsciously
    0.89
     instinctively
    0.88
     ages
    0.88
     perceive
    0.85
     psychologically
    0.85
     perce
    0.84
     enrolled
    0.84
     chronically
    0.82
     experiencing
    0.82
    Activations Density 0.260%

    No Known Activations