© 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. 19315
    Prev
    Next
    INDEX
    Explanations

    confessions or admissions of personal experiences and thoughts

    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
    447
    +0.10
    0.5%
    425
    +0.10
    0.5%
    92
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    431
    +0.10
    0.02
    718
    +0.10
    0.03
    425
    +0.10
    0.02
    Negative Logits
    DragonMagazine
    -0.74
    Liber
    -0.71
    ERT
    -0.70
    CLUS
    -0.68
    Friends
    -0.66
    icles
    -0.65
    Residents
    -0.65
    say
    -0.65
    AMI
    -0.64
    lake
    -0.64
    POSITIVE LOGITS
     mistakes
    1.20
     underestimated
    1.08
     underest
    1.03
     sometimes
    0.91
     shortcomings
    0.90
     mistake
    0.90
     lacked
    0.90
     regrets
    0.89
     errors
    0.88
     mishand
    0.88
    Activations Density 0.144%

    No Known Activations