© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 49152
    9. 8-RES_FS49152-JB
    11. 41763
    Prev
    Next
    INDEX
    Explanations

    numerical and contact information

    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
    726
    +0.10
    0.5%
    448
    +0.09
    0.4%
    380
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    380
    +0.10
    0.03
    726
    +0.09
    0.03
    258
    +0.09
    0.03
    Negative Logits
     stacking
    -0.79
     sweets
    -0.77
     clips
    -0.69
     footing
    -0.68
     losers
    -0.68
     stacked
    -0.67
     palate
    -0.67
     cull
    -0.64
     reused
    -0.64
     expended
    -0.63
    POSITIVE LOGITS
    Fax
    1.06
    email
    1.03
    Email
    1.01
    Phone
    1.01
     hotline
    0.96
    Tel
    0.93
    contact
    0.91
     tradem
    0.91
    76561
    0.90
    fax
    0.90
    Activations Density 0.099%

    No Known Activations