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

    Neuronpedia

    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 12288
    9. 8-RES_FS12288-JB
    11. 1792
    Prev
    Next
    INDEX
    Explanations

    descriptions of dislike, melancholy, and inner turmoil

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs12288-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    12,288
    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
    419
    +0.09
    0.4%
    237
    +0.09
    0.4%
    557
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    678
    +0.09
    0.06
    326
    +0.09
    0.07
    526
    +0.09
    0.00
    Negative Logits
    umbn
    -0.85
    Problem
    -0.79
    Fact
    -0.76
    ENTION
    -0.75
    =-=-=-=-
    -0.73
    CLASS
    -0.72
    ilater
    -0.72
    Assistant
    -0.71
    Myth
    -0.71
    Option
    -0.70
    POSITIVE LOGITS
     throb
    1.22
     shimmer
    1.17
     crimson
    1.10
     dripping
    1.07
     gle
    1.05
     stained
    1.03
     sparkling
    1.01
     ooz
    1.00
     streaks
    0.98
     haze
    0.97
    Activations Density 0.665%

    No Known Activations