© 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 24576
    9. 8-RES_FS24576-JB
    11. 7397
    Prev
    Next
    INDEX
    Explanations

    words and phrases related to government organizations and infrastructure projects

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs24576-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    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
    447
    +0.16
    0.7%
    625
    +0.11
    0.5%
    152
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    146
    +0.16
    0.06
    508
    +0.11
    0.04
    617
    +0.10
    0.04
    Negative Logits
    osphere
    -0.78
    ularity
    -0.78
    obin
    -0.75
     trilogy
    -0.75
    iera
    -0.70
     streak
    -0.69
     quot
    -0.67
     Sina
    -0.67
    vre
    -0.66
     Chomsky
    -0.66
    POSITIVE LOGITS
     officials
    1.30
     Commissioner
    1.16
     spokeswoman
    1.09
     Officials
    1.08
     spokesman
    1.08
     inspector
    1.06
     Inspector
    1.05
     Employees
    1.05
     inspectors
    1.05
     commissioners
    1.04
    Activations Density 0.216%

    No Known Activations