© 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 24576
    9. 8-RES_FS24576-JB
    11. 18855
    Prev
    Next
    INDEX
    Explanations

    verbs and phrases indicating action or implementation, especially related to policies or actions planned or taken

    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
    588
    +0.11
    0.5%
    197
    +0.10
    0.5%
    689
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    588
    +0.11
    0.03
    197
    +0.10
    0.02
    665
    +0.10
    0.03
    Negative Logits
    ibur
    -0.64
     Vaughan
    -0.62
    ajo
    -0.61
    paio
    -0.61
    á
    -0.58
    whatever
    -0.58
    article
    -0.58
    Scroll
    -0.56
    govtrack
    -0.56
    Thor
    -0.55
    POSITIVE LOGITS
     properly
    1.06
     correctly
    0.99
     incorrectly
    0.81
     individually
    0.79
     appropriately
    0.73
    ividually
    0.72
     improperly
    0.72
     toget
    0.67
     exceeds
    0.67
     aloud
    0.67
    Activations Density 0.083%

    No Known Activations