© 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. 24290
    Prev
    Next
    INDEX
    Explanations

    references to the concept of time, particularly indicating duration

    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
    400
    +0.11
    0.5%
    238
    +0.09
    0.5%
    529
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    93
    +0.11
    0.03
    251
    +0.09
    0.03
    493
    +0.09
    0.03
    Negative Logits
    arth
    -0.73
    itational
    -0.63
     flux
    -0.62
    asive
    -0.62
    olt
    -0.58
     Dating
    -0.57
    enda
    -0.56
    RPG
    -0.56
    icks
    -0.55
    CB
    -0.55
    POSITIVE LOGITS
     enough
    0.98
    etheless
    0.94
    itud
    0.80
    thing
    0.72
    fw
    0.72
     ago
    0.72
    enough
    0.71
    terday
    0.69
     outweigh
    0.66
    ishly
    0.65
    Activations Density 0.108%

    No Known Activations