© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 23947
    Prev
    Next
    INDEX
    Explanations

    the presence of the word "there" in various contexts

    oai_token-act-pair · gpt-4o-miniTriggered 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
    447
    +0.12
    0.5%
    314
    +0.11
    0.5%
    665
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    423
    +0.12
    0.01
    504
    +0.11
    0.01
    651
    +0.10
    0.01
    Negative Logits
    eers
    -0.70
    %%%%
    -0.66
    nr
    -0.63
    HH
    -0.62
    YES
    -0.62
    OTS
    -0.61
     Nationals
    -0.60
    sup
    -0.59
    rongh
    -0.59
    Ca
    -0.58
    POSITIVE LOGITS
     onwards
    1.26
     onward
    1.13
     sprang
    0.81
    annis
    0.78
     proceeded
    0.75
    abouts
    0.73
    upon
    0.72
    forward
    0.72
    forth
    0.71
     piv
    0.70
    Activations Density 0.016%

    No Known Activations