© 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. 9375
    Prev
    Next
    INDEX
    Explanations

    detailed narratives or stories related to past events or ongoing situations

    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
    393
    +0.11
    0.5%
    150
    +0.10
    0.5%
    106
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    170
    +0.11
    0.03
    106
    +0.10
    0.03
    150
    +0.10
    0.03
    Negative Logits
    anooga
    -0.78
    cn
    -0.73
    icator
    -0.72
    orah
    -0.70
    ombies
    -0.70
    ÙIJ
    -0.69
    د
    -0.67
     Cosponsors
    -0.67
    ctors
    -0.67
    named
    -0.67
    POSITIVE LOGITS
     uninterrupted
    1.12
     campaigning
    0.98
     silence
    0.91
     continuous
    0.89
     frustration
    0.89
     reckoning
    0.88
     stagnation
    0.88
     struggle
    0.87
     intense
    0.84
     exposure
    0.84
    Activations Density 0.113%

    No Known Activations