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 1536
    5. 8-RES_FS1536-JB
    6. 321
    Prev
    Next
    INDEX
    Explanations

    organizations, acronyms, and government-related terms

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs1536-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    1,536
    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
    627
    +0.12
    0.6%
    373
    +0.12
    0.5%
    364
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    627
    +0.12
    0.07
    659
    +0.12
    0.09
    314
    +0.11
    0.01
    Negative Logits
     depress
    -0.65
     weap
    -0.61
     piston
    -0.61
    ozyg
    -0.61
    gencies
    -0.59
    akespeare
    -0.58
     neigh
    -0.57
     illust
    -0.57
     caution
    -0.57
     hinge
    -0.57
    POSITIVE LOGITS
    )
    1.02
    ),
    1.01
    UA
    0.94
    )'
    0.93
    ISO
    0.88
    DS
    0.87
    )—
    0.86
    )[
    0.86
    DP
    0.85
    GP
    0.85
    Activations Density 1.290%

    No Known Activations