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 49152
    5. 8-RES_FS49152-JB
    6. 43235
    Prev
    Next
    INDEX
    Explanations

    phrases that indicate potential future events or outcomes

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs49152-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    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%
    718
    +0.12
    0.5%
    665
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    665
    +0.13
    0.05
    608
    +0.12
    0.04
    718
    +0.10
    0.04
    Negative Logits
    quickShipAvailable
    -0.80
    urance
    -0.76
    abiding
    -0.75
    raining
    -0.63
    _.
    -0.61
    regulation
    -0.61
    calling
    -0.61
    ivated
    -0.61
     Perspective
    -0.60
     Measures
    -0.59
    POSITIVE LOGITS
     participate
    1.37
     receive
    1.24
     partake
    1.15
     succeed
    1.15
     nominate
    1.13
     capitalize
    1.13
     join
    1.13
     accompany
    1.11
     marry
    1.10
    ilet
    1.04
    Activations Density 0.107%

    No Known Activations