© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    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. 18016
    Prev
    Next
    INDEX
    Explanations

    directive statements suggesting a manipulation of power and control over others

    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
    +0.15
    0.7%
    447
    +0.13
    0.6%
    640
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    6
    +0.15
    0.03
    9
    +0.13
    0.03
    176
    +0.12
    0.02
    Negative Logits
    orate
    -0.71
    rose
    -0.62
    alian
    -0.61
     bonded
    -0.59
     abl
    -0.58
    guard
    -0.58
     purse
    -0.58
     lad
    -0.58
     fres
    -0.58
    liner
    -0.58
    POSITIVE LOGITS
     Lastly
    1.31
     Finally
    1.21
     Whether
    1.07
     Ibid
    1.07
     Beware
    1.04
     Unable
    1.03
     Conclusion
    1.03
     Speaking
    1.01
     Occasionally
    1.01
     Regarding
    1.00
    Activations Density 0.085%

    No Known Activations