© 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 49152
    9. 8-RES_FS49152-JB
    11. 37229
    Prev
    Next
    INDEX
    Explanations

    references to measurements of distance

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs49152-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    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
    183
    +0.10
    0.4%
    756
    +0.09
    0.4%
    113
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    225
    +0.10
    0.01
    89
    +0.09
    0.01
    588
    +0.09
    0.01
    Negative Logits
    ulate
    -0.84
    stal
    -0.75
    essee
    -0.75
    umption
    -0.73
    steen
    -0.72
    ovy
    -0.71
    ovie
    -0.69
    udeb
    -0.69
    dom
    -0.68
    icious
    -0.67
    POSITIVE LOGITS
     travelled
    1.17
     traveled
    1.04
     distances
    1.04
     distance
    0.96
     separating
    0.84
    finder
    0.79
     between
    0.78
     Distance
    0.77
     flung
    0.74
    AGE
    0.74
    Activations Density 0.008%

    No Known Activations