© 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. 16113
    Prev
    Next
    INDEX
    Explanations

    information related to political negotiations, government powers, and public policy debates

    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
    +0.15
    0.7%
    102
    +0.11
    0.5%
    320
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    55
    +0.15
    0.08
    102
    +0.11
    0.07
    566
    +0.11
    0.06
    Negative Logits
    strous
    -0.66
     Might
    -0.65
     traverse
    -0.64
    çīĪ
    -0.62
     hail
    -0.62
    eday
    -0.61
     brace
    -0.61
     srfAttach
    -0.61
     wonders
    -0.60
    might
    -0.58
    POSITIVE LOGITS
     able
    1.21
     anymore
    1.15
     bothered
    1.09
     necessarily
    0.95
     counted
    0.94
     anywhere
    0.89
    icable
    0.87
     tolerated
    0.87
     harmed
    0.86
    fitting
    0.86
    Activations Density 0.134%

    No Known Activations