© 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. 4758
    Prev
    Next
    INDEX
    Explanations

    strong adverbs indicating continuation or opposition

    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
    373
    +0.13
    0.6%
    459
    +0.09
    0.4%
    468
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    468
    +0.13
    0.02
    760
    +0.09
    0.02
    85
    +0.09
    0.02
    Negative Logits
    duct
    -0.69
     è£ıç
    -0.67
    alties
    -0.64
    ASE
    -0.63
    orum
    -0.62
    rys
    -0.62
    arthed
    -0.60
    ranch
    -0.60
    ospace
    -0.59
    zag
    -0.59
    POSITIVE LOGITS
    birth
    0.90
     unanswered
    0.78
     undecided
    0.73
     reeling
    0.72
    adays
    0.72
     unresolved
    0.72
     unclear
    0.71
    noon
    0.67
     unsure
    0.67
     clinging
    0.66
    Activations Density 0.026%

    No Known Activations