Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 1536
    5. 8-RES_FS1536-JB
    6. 454
    Prev
    Next
    INDEX
    Explanations

    references to political figures, decisions, or controversies

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs1536-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    1,536
    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
    139
    +0.10
    0.5%
    343
    +0.09
    0.4%
    203
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    139
    +0.10
    0.11
    443
    +0.09
    0.07
    741
    +0.09
    0.08
    Negative Logits
     mats
    -0.87
     NAD
    -0.81
    AGES
    -0.73
     metre
    -0.73
     Sensor
    -0.72
     convent
    -0.72
    atos
    -0.71
     Mages
    -0.71
     overload
    -0.71
     Sisters
    -0.71
    POSITIVE LOGITS
     pard
    1.15
     veto
    1.14
    realDonaldTrump
    1.13
     appoint
    1.11
     presidential
    1.09
     aides
    1.08
    Hillary
    1.07
     pardon
    1.07
     presidency
    1.05
    Obama
    1.05
    Activations Density 5.441%

    No Known Activations