© 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 EvalsExportsinterp-engineNEWAPI 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. 14315
    Prev
    Next
    INDEX
    Explanations

    names of individuals and locations followed by actions or statements attributed to them

    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
    55
    +0.13
    0.6%
    196
    +0.12
    0.5%
    564
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    443
    +0.13
    0.05
    139
    +0.12
    0.06
    55
    +0.10
    0.07
    Negative Logits
     selves
    -0.85
    rift
    -0.69
     sleeper
    -0.68
    selves
    -0.67
    rossover
    -0.65
    acters
    -0.65
    Jews
    -0.64
     analytic
    -0.64
     sects
    -0.64
     penetration
    -0.64
    POSITIVE LOGITS
     appoint
    1.18
     vetoed
    1.17
     veto
    1.16
     presided
    1.00
     appointed
    0.97
     announced
    0.95
     briefed
    0.95
     oversaw
    0.91
     oversee
    0.91
     announce
    0.90
    Activations Density 0.614%

    No Known Activations