© 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. 9325
    Prev
    Next
    INDEX
    Explanations

    unacceptable behavior in different scenarios or settings

    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
    481
    +0.14
    0.6%
    674
    +0.14
    0.6%
    373
    +0.13
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    674
    +0.14
    0.03
    119
    +0.14
    0.03
    101
    +0.13
    0.06
    Negative Logits
     nonetheless
    -0.81
     theirs
    -0.78
    .</
    -0.72
     accordingly
    -0.67
    Indeed
    -0.66
     nevertheless
    -0.65
    Certainly
    -0.60
    Details
    -0.57
    Developer
    -0.57
    KNOWN
    -0.57
    POSITIVE LOGITS
    etooth
    0.61
    £ı
    0.60
     divor
    0.59
     Quit
    0.58
     Assass
    0.58
    */(
    0.57
    ulhu
    0.57
     Voters
    0.56
     Airl
    0.56
     Smoking
    0.55
    Activations Density 0.730%

    No Known Activations