© 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. 1761
    Prev
    Next
    INDEX
    Explanations

    phrases related to instructions or steps

    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
    36
    +0.26
    1.2%
    640
    +0.11
    0.5%
    362
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    36
    +0.26
    0.09
    250
    +0.11
    0.03
    362
    +0.10
    0.04
    Negative Logits
    upon
    -0.83
    Events
    -0.68
    ties
    -0.67
    lords
    -0.66
    oons
    -0.65
    Contents
    -0.65
    overs
    -0.64
    otle
    -0.63
    ibraries
    -0.62
    aurus
    -0.61
    POSITIVE LOGITS
     whim
    1.16
     basis
    1.10
     dime
    1.07
     behalf
    1.00
     leash
    0.98
     treadmill
    0.98
     rainy
    0.94
     doorstep
    0.94
     occasions
    0.94
     scale
    0.93
    Activations Density 0.059%

    No Known Activations