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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsBlogSlackPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 3072
    5. 8-RES_FS3072-JB
    6. 1705
    Prev
    Next
    INDEX
    Explanations

    mentions of books or articles and their authors

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs3072-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    3,072
    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
    760
    +0.12
    0.5%
    557
    +0.10
    0.5%
    18
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    680
    +0.12
    0.07
    175
    +0.10
    0.06
    498
    +0.10
    0.09
    Negative Logits
     customs
    -0.83
     Customs
    -0.68
     broom
    -0.68
     crane
    -0.67
    animate
    -0.67
     wink
    -0.67
     ambulance
    -0.66
     valve
    -0.66
     whim
    -0.66
    bucks
    -0.65
    POSITIVE LOGITS
     essays
    1.41
     articles
    1.16
     essay
    1.15
    published
    1.12
     excerpts
    1.09
     insightful
    1.07
     blogs
    1.07
    pdf
    1.06
     reprinted
    1.05
     Understanding
    1.05
    Activations Density 2.254%

    No Known Activations