© 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 EvalsExportsinterp-engineNEWAPI 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. 13268
    Prev
    Next
    INDEX
    Explanations

    phrases related to controversies or investigations

    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
    114
    +0.13
    0.6%
    556
    +0.11
    0.5%
    428
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    114
    +0.13
    0.05
    195
    +0.11
    0.02
    291
    +0.09
    0.05
    Negative Logits
     actionGroup
    -0.94
    ktop
    -0.84
     Mobility
    -0.76
     Container
    -0.72
    zie
    -0.72
    ourning
    -0.71
     mobility
    -0.68
    enfranch
    -0.68
    isites
    -0.68
    sembly
    -0.66
    POSITIVE LOGITS
     assertions
    1.43
     debunked
    1.35
     untrue
    1.33
     claims
    1.26
     falsehood
    1.22
     allegations
    1.21
     factual
    1.20
     inaccurate
    1.20
     refuted
    1.18
     debunk
    1.18
    Activations Density 0.624%

    No Known Activations