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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 24576
    5. 8-RES_FS24576-JB
    6. 7172
    Prev
    Next
    INDEX
    Explanations

    names related to individuals involved in national security and classified information leaks

    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%
    108
    +0.10
    0.5%
    333
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    346
    +0.15
    0.04
    386
    +0.10
    0.03
    129
    +0.09
    0.02
    Negative Logits
    entary
    -0.83
    bowl
    -0.78
    sembly
    -0.76
     Farmers
    -0.74
     Hort
    -0.74
     Agric
    -0.73
     Hitchcock
    -0.73
     Veter
    -0.71
    urion
    -0.71
     Governors
    -0.71
    POSITIVE LOGITS
     WikiLeaks
    1.27
     Wikileaks
    1.22
     whistlebl
    1.21
     revelations
    1.15
    ileaks
    1.11
     disclosures
    1.11
     pseudonym
    1.11
     hacked
    1.09
     whistleblower
    1.05
    Leaks
    1.04
    Activations Density 0.377%

    No Known Activations