© 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. 637
    Prev
    Next
    INDEX
    Explanations

    phrases related to international political conflicts and events

    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
    154
    +0.14
    0.6%
    745
    +0.10
    0.4%
    385
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    154
    +0.14
    0.16
    494
    +0.10
    0.14
    725
    +0.09
    0.11
    Negative Logits
     Whitney
    -0.86
     Cary
    -0.86
     Lifetime
    -0.85
    Consumer
    -0.84
     Patent
    -0.82
     AOL
    -0.77
     Edison
    -0.77
     Oaks
    -0.76
    Customer
    -0.75
     Barney
    -0.75
    POSITIVE LOGITS
     militants
    1.48
     Islamist
    1.45
     militias
    1.41
     rebels
    1.41
     Houth
    1.40
     insurgents
    1.39
     separatist
    1.38
     separatists
    1.38
     Hezbollah
    1.36
     militant
    1.36
    Activations Density 8.324%

    No Known Activations