© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 768
    9. 8-RES_FS768-JB
    11. 708
    Prev
    Next
    INDEX
    Explanations

    phrases related to medical procedures and healthcare policies

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs768-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    768
    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
    447
    +0.16
    0.8%
    481
    +0.15
    0.7%
    529
    +0.12
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    529
    +0.16
    0.13
    665
    +0.15
    0.17
    608
    +0.12
    0.19
    Negative Logits
     partName
    -0.81
     disappoint
    -0.69
    Eva
    -0.68
    fortunately
    -0.68
     moments
    -0.68
     misfortune
    -0.67
    Sym
    -0.67
     sorrow
    -0.66
     scratches
    -0.64
     exclaim
    -0.64
    POSITIVE LOGITS
     compulsory
    1.00
     centralized
    0.96
     decentral
    0.93
     interoper
    0.92
     privatization
    0.91
     infrastructure
    0.90
     mandatory
    0.88
     abol
    0.88
     privat
    0.88
     universal
    0.87
    Activations Density 8.756%

    No Known Activations