© 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. 17307
    Prev
    Next
    INDEX
    Explanations

    organizations or specific entities, particularly related to the public sector

    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
    169
    +0.10
    0.4%
    288
    +0.10
    0.4%
    179
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    700
    +0.10
    0.03
    527
    +0.10
    0.02
    303
    +0.10
    0.02
    Negative Logits
    kson
    -0.78
    ansas
    -0.74
    orsi
    -0.72
    oldemort
    -0.70
    xual
    -0.70
    Scroll
    -0.69
    ikawa
    -0.69
    Downloadha
    -0.67
    netflix
    -0.66
    innon
    -0.65
    POSITIVE LOGITS
     Relations
    0.94
     Affairs
    0.91
     Opinion
    0.88
     Interest
    0.86
     Accountability
    0.86
     Safety
    0.84
     Service
    0.83
     Integrity
    0.81
     Schools
    0.81
    ity
    0.80
    Activations Density 0.021%

    No Known Activations