© 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 EvalsExportsinterp-engineNEWAPI 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. 22588
    Prev
    Next
    INDEX
    Explanations

    negative sentiments or criticisms about various subjects

    oai_token-act-pair · gpt-4o-miniTriggered 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
    380
    +0.11
    0.5%
    544
    +0.10
    0.4%
    273
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    43
    +0.11
    0.08
    546
    +0.10
    0.08
    720
    +0.10
    0.08
    Negative Logits
     guiActiveUnfocused
    -0.81
    boro
    -0.69
    STATE
    -0.69
    Includes
    -0.67
    ļéĨĴ
    -0.67
    ../
    -0.66
    Accessory
    -0.65
    resents
    -0.64
    interstitial
    -0.64
    handled
    -0.62
    POSITIVE LOGITS
     however
    1.77
     meanwhile
    1.41
     though
    1.38
     moreover
    1.37
     incidentally
    1.04
     alas
    1.02
     therefore
    1.01
     anyway
    1.00
     indeed
    0.92
     meantime
    0.90
    Activations Density 0.480%

    No Known Activations