© 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. 7228
    Prev
    Next
    INDEX
    Explanations

    guidelines related to content strategy and audience engagement

    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
    168
    +0.10
    0.4%
    186
    +0.10
    0.4%
    653
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    762
    +0.10
    0.08
    224
    +0.10
    0.07
    401
    +0.09
    0.06
    Negative Logits
    inventoryQuantity
    -0.71
     Canaver
    -0.70
     Lug
    -0.67
     Deadly
    -0.62
     Stre
    -0.60
    neck
    -0.60
    ohyd
    -0.60
    ahime
    -0.59
    Fla
    -0.59
    ivist
    -0.59
    POSITIVE LOGITS
     reacted
    1.01
     couldn
    1.01
     wouldn
    1.01
     dare
    0.99
     alike
    0.99
     would
    0.98
     recognise
    0.97
     knew
    0.97
     perceive
    0.94
     will
    0.93
    Activations Density 0.443%

    No Known Activations