© 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. 18009
    Prev
    Next
    INDEX
    Explanations

    phrases related to physical sensations or actions involving physical objects

    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
    156
    +0.10
    0.5%
    503
    +0.10
    0.4%
    659
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    125
    +0.10
    0.01
    232
    +0.10
    0.01
    503
    +0.10
    0.01
    Negative Logits
     reconciliation
    -0.68
     neglect
    -0.67
     dating
    -0.67
     Luther
    -0.67
     masculinity
    -0.64
     Archdemon
    -0.61
     rested
    -0.60
    WAYS
    -0.60
    urally
    -0.59
    geist
    -0.58
    POSITIVE LOGITS
    irts
    1.23
    atters
    1.20
    ishable
    1.14
    awk
    1.14
    irting
    1.13
    irrel
    1.12
    alid
    1.11
    ount
    1.09
    atter
    1.08
    atted
    1.08
    Activations Density 0.020%

    No Known Activations