© 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 12288
    9. 8-RES_FS12288-JB
    11. 10189
    Prev
    Next
    INDEX
    Explanations

    questions related to ethical and social issues

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs12288-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    12,288
    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
    285
    +0.16
    0.7%
    166
    +0.10
    0.4%
    6
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    285
    +0.16
    0.07
    6
    +0.10
    0.08
    232
    +0.10
    0.05
    Negative Logits
     arm
    -0.71
     corrid
    -0.69
    ikuman
    -0.66
    elist
    -0.66
     yak
    -0.65
     dex
    -0.64
    eared
    -0.63
     booked
    -0.62
    cipled
    -0.61
    ishable
    -0.61
    POSITIVE LOGITS
     Surely
    1.32
     Wouldn
    1.27
     Nope
    1.27
     Answer
    1.26
     Certainly
    1.22
    Answer
    1.19
     Probably
    1.17
     Would
    1.14
     Well
    1.14
     Isn
    1.11
    Activations Density 0.122%

    No Known Activations