© 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 24576
    9. 8-RES_FS24576-JB
    11. 8653
    Prev
    Next
    INDEX
    Explanations

    mentions of different types of holds, such as hair styling products, sports maneuvers, and legal terms

    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
    463
    +0.11
    0.5%
    183
    +0.11
    0.5%
    540
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    531
    +0.11
    0.03
    582
    +0.11
    0.02
    463
    +0.10
    0.03
    Negative Logits
    schild
    -0.70
    icycle
    -0.67
     Daylight
    -0.66
    ople
    -0.66
    tnc
    -0.65
    iscal
    -0.65
    issance
    -0.65
    gnu
    -0.65
    éĹ
    -0.64
    chrom
    -0.63
    POSITIVE LOGITS
    hold
    1.30
    erness
    1.10
    holder
    1.07
    holders
    1.07
    holding
    1.01
    down
    0.90
    outs
    0.87
     hold
    0.87
    overs
    0.85
    fast
    0.83
    Activations Density 0.037%

    No Known Activations