Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 24576
    5. 8-RES_FS24576-JB
    6. 21805
    Prev
    Next
    INDEX
    Explanations

    details about violent incidents involving people or casualties

    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
    447
    +0.14
    0.6%
    241
    +0.13
    0.6%
    47
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    241
    +0.14
    0.04
    516
    +0.13
    0.04
    47
    +0.10
    0.03
    Negative Logits
    idth
    -0.80
    Completed
    -0.68
    Measure
    -0.67
    uncture
    -0.66
    wcsstore
    -0.66
     compulsion
    -0.65
    Limited
    -0.64
    Consumer
    -0.61
    IELD
    -0.61
     disclaimer
    -0.61
    POSITIVE LOGITS
     fleeing
    1.02
     unarmed
    0.99
     indiscrim
    0.86
     injuring
    0.84
     innocent
    0.83
     innoc
    0.83
     wounding
    0.82
     civilians
    0.82
     smugglers
    0.81
     bystanders
    0.81
    Activations Density 0.109%

    No Known Activations