© 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. 6611
    Prev
    Next
    INDEX
    Explanations

    contrasting clauses or comparisons

    oai_token-act-pair · gpt-3.5-turboTriggered 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
    481
    +0.18
    0.8%
    579
    +0.11
    0.5%
    512
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    512
    +0.18
    0.06
    36
    +0.11
    0.06
    393
    +0.10
    0.06
    Negative Logits
    eton
    -0.92
    aido
    -0.83
    votes
    -0.76
     winters
    -0.74
    20439
    -0.73
    laus
    -0.68
    oren
    -0.68
    lees
    -0.68
    count
    -0.67
    tones
    -0.66
    POSITIVE LOGITS
     conduit
    1.14
     tool
    1.07
     shorthand
    0.99
     gateway
    0.97
     manifestation
    0.95
     condu
    0.95
     indispensable
    0.95
     cornerstone
    0.93
     messenger
    0.91
     symbol
    0.91
    Activations Density 0.365%

    No Known Activations