© 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 12288
    9. 8-RES_FS12288-JB
    11. 681
    Prev
    Next
    INDEX
    Explanations

    references to being a part of a group or community

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs12288-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    12,288
    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
    373
    +0.13
    0.6%
    676
    +0.12
    0.5%
    618
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    676
    +0.13
    0.05
    618
    +0.12
    0.04
    597
    +0.10
    0.04
    Negative Logits
    culosis
    -0.76
     ceilings
    -0.61
    stasy
    -0.59
     heights
    -0.59
    olesc
    -0.59
     calves
    -0.56
    izu
    -0.56
     briefs
    -0.55
    ITH
    -0.53
    ricanes
    -0.52
    POSITIVE LOGITS
    aking
    1.00
    ridge
    0.92
    heartedly
    0.84
    meal
    0.82
    nered
    0.82
    icularly
    0.80
    ially
    0.78
    ners
    0.75
    dated
    0.74
    icipated
    0.73
    Activations Density 0.057%

    No Known Activations