© 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. 717
    Prev
    Next
    INDEX
    Explanations

    sentence structures with a negative context or controversy

    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
    640
    +0.28
    1.3%
    649
    +0.12
    0.6%
    373
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    640
    +0.28
    0.30
    94
    +0.12
    0.23
    539
    +0.12
    0.16
    Negative Logits
    heit
    -0.71
    lections
    -0.69
    boro
    -0.65
    utical
    -0.64
    minist
    -0.64
     endeav
    -0.60
     hydra
    -0.57
    burn
    -0.57
    mbuds
    -0.57
     endeavour
    -0.56
    POSITIVE LOGITS
    Scroll
    1.04
    PHOTOS
    0.96
    Asked
    0.83
    However
    0.82
     Related
    0.81
    Bus
    0.79
    Reporting
    0.78
    SEE
    0.77
    Meanwhile
    0.76
    Authorities
    0.76
    Activations Density 1.688%

    No Known Activations