© 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. 3104
    Prev
    Next
    INDEX
    Explanations

    phrases related to losses and changes

    oai_token-act-pair · gpt-3.5-turbo
    Triggered 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
    Neuron Alignment
    Index
    Value
    % of L₁
    245
    +0.13
    0.6%
    137
    +0.11
    0.5%
    763
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    245
    +0.13
    0.03
    761
    +0.11
    0.03
    223
    +0.09
    0.03
    Negative Logits
    SPONSORED
    -0.85
    DIT
    -0.70
    LLOW
    -0.70
    INFO
    -0.70
    WARN
    -0.69
    DEBUG
    -0.67
    soDeliveryDate
    -0.64
    EGIN
    -0.62
    CHECK
    -0.62
    EEP
    -0.61
    POSITIVE LOGITS
     altogether
    1.00
     virginity
    0.95
     :(
    0.94
     due
    0.89
     prematurely
    0.78
     dignity
    0.77
     credibility
    0.75
     inhib
    0.73
     freedoms
    0.73
     momentum
    0.73
    Activations Density 0.221%

    No Known Activations