© 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. 15792
    Prev
    Next
    INDEX
    Explanations

    phrases emphasizing the degree of something

    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
    760
    +0.10
    0.4%
    438
    +0.09
    0.4%
    297
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    760
    +0.10
    0.03
    297
    +0.09
    0.03
    438
    +0.09
    0.03
    Negative Logits
     DRAG
    -0.86
    elsen
    -0.81
    venant
    -0.79
    ULTS
    -0.78
    ulkan
    -0.76
    PLIED
    -0.68
    erson
    -0.67
    ults
    -0.66
    adium
    -0.65
    ipers
    -0.65
    POSITIVE LOGITS
     frankly
    0.99
     literally
    0.76
     alright
    0.75
     nice
    0.72
     possibly
    0.70
     nifty
    0.70
     tame
    0.68
     scary
    0.67
     enough
    0.65
     chilly
    0.65
    Activations Density 0.017%

    No Known Activations