© 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 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. 7865
    Prev
    Next
    INDEX
    Explanations

    references to the concept of "veto" or instances of the word "veto" with different variations

    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
    500
    +0.10
    0.5%
    218
    +0.10
    0.4%
    78
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    500
    +0.10
    0.01
    78
    +0.10
    0.01
    286
    +0.09
    0.01
    Negative Logits
    ndra
    -0.85
    asma
    -0.75
    nsic
    -0.73
    geist
    -0.73
    istics
    -0.72
    inav
    -0.71
    negie
    -0.71
    ibaba
    -0.70
    adr
    -0.69
    arers
    -0.69
    POSITIVE LOGITS
     veto
    1.10
     vetoed
    1.01
    zai
    0.81
     amnesty
    0.77
     override
    0.76
     bluff
    0.72
     Feinstein
    0.71
     ratification
    0.70
     decree
    0.70
     overr
    0.69
    Activations Density 0.025%

    No Known Activations