© 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 49152
    9. 8-RES_FS49152-JB
    11. 25757
    Prev
    Next
    INDEX
    Explanations

    chemical compounds and their interactions in scientific contexts

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs49152-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    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
    480
    +0.13
    1.4%
    361
    +0.05
    0.5%
    11
    +0.04
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    620
    +0.13
    -0.01
    283
    +0.05
    -0.01
    469
    +0.04
    0.01
    Negative Logits
     fucking
    -0.27
     legal
    -0.26
     intellectual
    -0.24
     fuckin
    -0.24
     techn
    -0.23
     anthrop
    -0.23
     copyright
    -0.23
     Legal
    -0.23
     foreign
    -0.23
     glor
    -0.23
    POSITIVE LOGITS
    inventoryQuantity
    0.33
    finding
    0.31
    inos
    0.29
     scrimmage
    0.29
    ieu
    0.29
    dayName
    0.28
    riott
    0.27
    yles
    0.25
    ggles
    0.25
    anza
    0.25
    Activations Density 2.050%

    No Known Activations