© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 20062
    Prev
    Next
    INDEX
    Explanations

    phrases related to negative weather conditions and disasters

    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
    551
    +0.15
    0.7%
    718
    +0.12
    0.5%
    49
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    551
    +0.15
    0.06
    49
    +0.12
    0.06
    24
    +0.10
    0.04
    Negative Logits
     Emerging
    -0.69
     Nort
    -0.69
    Recommended
    -0.67
     Guilty
    -0.66
     Emir
    -0.66
     Rust
    -0.65
     Blueprint
    -0.64
     Ideal
    -0.64
     Vessel
    -0.63
     Lambert
    -0.61
    POSITIVE LOGITS
     endure
    1.15
     rely
    1.14
     contend
    1.06
     fend
    1.05
     wait
    1.04
     overcome
    1.01
     intervene
    1.00
     cope
    0.99
    wered
    0.99
     undergo
    0.99
    Activations Density 0.099%

    No Known Activations