© 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 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. 10377
    Prev
    Next
    INDEX
    Explanations

    phrases related to agreements, disagreements, and political or legal issues

    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
    +0.26
    1.2%
    256
    +0.11
    0.5%
    27
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    86
    +0.26
    0.05
    762
    +0.11
    0.06
    718
    +0.11
    0.05
    Negative Logits
     Miracle
    -0.78
    advertisement
    -0.74
    Atl
    -0.73
    stuff
    -0.72
     Atlantis
    -0.70
    Delivery
    -0.65
     Saga
    -0.64
    duration
    -0.63
     Explosion
    -0.63
     Rout
    -0.63
    POSITIVE LOGITS
     endorse
    1.22
     vehemently
    1.19
     preferring
    1.14
     sympath
    1.13
     opposes
    1.12
     dislike
    1.12
     disdain
    1.12
     urge
    1.11
     condemn
    1.10
     disappro
    1.09
    Activations Density 0.507%

    No Known Activations