© 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. 6824
    Prev
    Next
    INDEX
    Explanations

    words and phrases related to political news and announcements

    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
    373
    +0.12
    0.5%
    314
    +0.11
    0.5%
    308
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    308
    +0.12
    0.02
    6
    +0.11
    0.02
    739
    +0.10
    0.01
    Negative Logits
     scope
    -0.69
    interstitial
    -0.69
    adden
    -0.67
    agate
    -0.67
    wagen
    -0.66
     servicing
    -0.65
    teenth
    -0.64
     appropri
    -0.64
    deals
    -0.62
    ouls
    -0.62
    POSITIVE LOGITS
    =-=-
    0.94
     Advertisement
    0.91
    =-=-=-=-
    0.90
     Comments
    0.80
    ĸļ
    0.79
    ADVERTISEMENT
    0.79
    =-=-=-=-=-=-=-=-
    0.79
     Ibid
    0.78
    kat
    0.77
    RANT
    0.76
    Activations Density 0.035%

    No Known Activations