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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsBlogSlackPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 768
    5. 8-RES_FS768-JB
    6. 358
    Prev
    Next
    INDEX
    Explanations

    the word "as" at the start of sentences

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs768-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    768
    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
    192
    +0.11
    0.5%
    393
    +0.11
    0.5%
    172
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    172
    +0.11
    0.13
    647
    +0.11
    0.12
    441
    +0.11
    0.11
    Negative Logits
    ople
    -0.65
    oller
    -0.63
     Import
    -0.59
    ãģł
    -0.58
    osc
    -0.58
    OUS
    -0.58
    Rub
    -0.56
    eeee
    -0.56
    BACK
    -0.56
    itiveness
    -0.55
    POSITIVE LOGITS
    pired
    1.24
     opposed
    1.23
    pires
    1.19
     evidenced
    1.15
    piring
    1.15
    ylum
    1.14
     well
    1.14
    pects
    1.12
    bestos
    1.08
    ynchron
    1.07
    Activations Density 1.818%

    No Known Activations