© 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 49152
    9. 8-RES_FS49152-JB
    11. 41481
    Prev
    Next
    INDEX
    Explanations

    instances of the word "upset," indicating emotions of dissatisfaction or distress

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs49152-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    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
    49
    +0.12
    0.6%
    267
    +0.12
    0.6%
    288
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    49
    +0.12
    0.01
    271
    +0.12
    0.01
    195
    +0.11
    0.01
    Negative Logits
    glas
    -0.85
     livest
    -0.81
    ographies
    -0.75
    audi
    -0.70
    istered
    -0.70
    gravity
    -0.69
    haar
    -0.69
    perty
    -0.69
    estones
    -0.68
    atures
    -0.67
    POSITIVE LOGITS
    dy
    0.86
    der
    0.80
     stomach
    0.75
    ingly
    0.74
    quished
    0.72
     upset
    0.72
    NESS
    0.70
    bur
    0.70
     roy
    0.69
    ful
    0.67
    Activations Density 0.008%

    No Known Activations