© 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 49152
    9. 8-RES_FS49152-JB
    11. 41756
    Prev
    Next
    INDEX
    Explanations

    references to news agencies

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs49152-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    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
    594
    +0.10
    0.5%
    373
    +0.10
    0.5%
    493
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    28
    +0.10
    0.03
    586
    +0.10
    0.02
    433
    +0.10
    0.02
    Negative Logits
    ãĢİ
    -0.74
     seminal
    -0.68
     langu
    -0.68
    gone
    -0.66
     chilling
    -0.64
    equality
    -0.62
    ¶
    -0.60
    judicial
    -0.58
     embod
    -0.57
    ãĢIJ
    -0.57
    POSITIVE LOGITS
    UTERS
    0.92
     Photo
    0.88
    CBC
    0.86
    /)
    0.86
    )
    0.86
    )"
    0.84
    +)
    0.80
    )."
    0.79
    )—
    0.79
    photo
    0.79
    Activations Density 0.019%

    No Known Activations