© 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 12288
    9. 8-RES_FS12288-JB
    11. 10629
    Prev
    Next
    INDEX
    Explanations

    references to legal or political concepts, specifically related to laws, policies, or government actions

    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
    481
    +0.34
    1.6%
    278
    +0.10
    0.5%
    154
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    762
    +0.34
    0.11
    435
    +0.10
    0.05
    481
    +0.10
    0.10
    Negative Logits
     precaution
    -0.66
    BuyableInstoreAndOnline
    -0.61
     backdrop
    -0.58
    \":
    -0.56
    idth
    -0.55
    pired
    -0.55
     antiqu
    -0.53
    Ku
    -0.52
    ãĤ«
    -0.52
     accus
    -0.51
    POSITIVE LOGITS
     anytime
    1.22
     WITHOUT
    1.02
     someday
    1.01
     without
    1.01
     sooner
    0.98
     depending
    0.96
     quicker
    0.88
     if
    0.87
     unless
    0.85
    depending
    0.83
    Activations Density 1.354%

    No Known Activations