© 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. 3227
    Prev
    Next
    INDEX
    Explanations

    phrases related to cyberattacks and computer security vulnerabilities

    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
    547
    +0.13
    0.6%
    195
    +0.10
    0.5%
    165
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    211
    +0.13
    0.01
    655
    +0.10
    0.03
    1
    +0.10
    0.03
    Negative Logits
     Connie
    -1.00
    chairs
    -0.87
    comfort
    -0.86
     canoe
    -0.86
    heny
    -0.82
    apolis
    -0.80
     apprentices
    -0.79
     Negro
    -0.79
    rentices
    -0.79
    Commission
    -0.79
    POSITIVE LOGITS
     malware
    1.89
     ransomware
    1.67
     hackers
    1.55
     malicious
    1.50
     hacker
    1.39
     attackers
    1.33
     encrypted
    1.29
     hacked
    1.28
     vulnerabilities
    1.28
     CVE
    1.27
    Activations Density 0.618%

    No Known Activations