© 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 768
    9. 8-RES_FS768-JB
    11. 734
    Prev
    Next
    INDEX
    Explanations

    instances of the word "of" occurring with emphasis, particularly in phrases or contexts that convey importance or strong impact

    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
    393
    +0.12
    0.5%
    526
    +0.11
    0.5%
    756
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    611
    +0.12
    0.12
    311
    +0.11
    0.04
    362
    +0.10
    0.13
    Negative Logits
     reperto
    -0.70
    uyomi
    -0.62
    Downloadha
    -0.60
    ynes
    -0.59
    Enlarge
    -0.59
     partName
    -0.58
     evaluations
    -0.57
     portray
    -0.57
    SpaceEngineers
    -0.56
    itent
    -0.56
    POSITIVE LOGITS
     course
    1.55
     sorts
    1.37
    course
    1.27
    icial
    1.10
    ortunately
    1.03
     ours
    1.03
     theirs
    1.02
     yours
    0.96
     hers
    0.95
     Course
    0.91
    Activations Density 2.490%

    No Known Activations