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 768
    5. 8-RES_FS768-JB
    6. 127
    Prev
    Next
    INDEX
    Explanations

    references to financial transactions or payments

    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
    366
    +0.09
    0.4%
    639
    +0.09
    0.4%
    654
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    654
    +0.09
    0.08
    35
    +0.09
    0.08
    352
    +0.09
    0.09
    Negative Logits
    ierre
    -0.79
    bum
    -0.71
    ocent
    -0.70
    ullivan
    -0.70
    urst
    -0.67
    ãĤ¼ãĤ¦ãĤ¹
    -0.67
    ocamp
    -0.66
    fw
    -0.66
    ################################
    -0.65
    swick
    -0.65
    POSITIVE LOGITS
     dues
    1.11
     taxes
    1.09
    checks
    1.07
     royalties
    1.07
     dividends
    1.03
     dearly
    1.02
     premiums
    1.01
    paying
    0.98
     homage
    0.97
     payable
    0.96
    Activations Density 3.098%

    No Known Activations