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

    Neuronpedia

    Jacobian LensNEW
    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. 12688
    Prev
    Next
    INDEX
    Explanations

    I'm sorry but I don't have enough information to determine what the neuron is looking for based on the provided activations

    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
    102
    +0.13
    0.6%
    7
    +0.12
    0.6%
    55
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    55
    +0.13
    0.06
    102
    +0.12
    0.06
    7
    +0.11
    0.05
    Negative Logits
    rones
    -0.78
    ĸļ
    -0.74
    enary
    -0.71
    enaries
    -0.68
    shall
    -0.67
    anwhile
    -0.66
    ulously
    -0.66
    ablishment
    -0.65
    strous
    -0.64
    ãĥ¼ãĥ«
    -0.64
    POSITIVE LOGITS
     able
    1.17
     tempted
    1.12
     forgiven
    1.05
     construed
    1.05
     considered
    1.04
     swayed
    0.98
     regarded
    0.96
     mistaken
    0.95
     viewed
    0.93
    acons
    0.87
    Activations Density 0.121%

    No Known Activations