© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW 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. 3116
    Prev
    Next
    INDEX
    Explanations

    references to a specific newspaper named "The Star" in different articles

    oai_token-act-pair · gpt-3.5-turbo
    Triggered 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
    Neuron Alignment
    Index
    Value
    % of L₁
    459
    +0.16
    0.7%
    339
    +0.13
    0.6%
    261
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    386
    +0.16
    0.02
    261
    +0.13
    0.02
    606
    +0.12
    0.02
    Negative Logits
    ĸļ
    -0.77
    Downloadha
    -0.75
    ĵĺ
    -0.73
    ecause
    -0.72
    yout
    -0.71
    odcast
    -0.71
    essee
    -0.71
    ¿½
    -0.70
    urally
    -0.68
     LIA
    -0.68
    POSITIVE LOGITS
    vation
    1.21
    bucks
    1.05
    ring
    1.01
    ved
    0.99
    fish
    0.96
    burst
    0.96
    light
    0.95
    ving
    0.85
    lite
    0.85
    lings
    0.84
    Activations Density 0.019%

    No Known Activations