© 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. 14801
    Prev
    Next
    INDEX
    Explanations

    phrases related to reports, news, and statements

    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
    481
    +0.14
    0.7%
    430
    +0.14
    0.6%
    234
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    234
    +0.14
    0.06
    563
    +0.14
    0.06
    452
    +0.12
    0.05
    Negative Logits
     srfAttach
    -0.69
     ILCS
    -0.68
    ardless
    -0.66
     unic
    -0.64
    ventus
    -0.64
    âĹ¼
    -0.63
    ntil
    -0.62
    TPPStreamerBot
    -0.61
     actionGroup
    -0.60
    Stock
    -0.60
    POSITIVE LOGITS
     concludes
    1.07
     adds
    0.98
     cites
    0.96
     alleges
    0.95
     contains
    0.94
     says
    0.93
     summarizes
    0.92
     emphasizes
    0.92
     includes
    0.91
     cited
    0.91
    Activations Density 0.165%

    No Known Activations