© 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. 9787
    Prev
    Next
    INDEX
    Explanations

    instances where the text expresses a strong positive or negative sentiment towards a product or experience

    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
    718
    +0.13
    0.6%
    251
    +0.11
    0.5%
    122
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    551
    +0.13
    0.04
    718
    +0.11
    0.04
    122
    +0.10
    0.04
    Negative Logits
    senal
    -0.69
     Appears
    -0.68
     Actions
    -0.67
    ILA
    -0.65
    artifacts
    -0.61
     Transparency
    -0.61
    interstitial
    -0.61
     invis
    -0.60
    farious
    -0.60
    oration
    -0.59
    POSITIVE LOGITS
     afford
    1.14
     muster
    0.93
     cope
    0.92
     foresee
    0.91
     convince
    0.89
     overcome
    0.89
     conceive
    0.89
     locate
    0.87
     persuade
    0.85
     pinpoint
    0.84
    Activations Density 0.071%

    No Known Activations