© 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. Apollo Research · Taylor · Identifying Functionally Important Features with End-to-End Sparse Dictionary Learning
    5. GPT2-Small
    7. Residual Similar CE E2E Recon
    9. 6-RES_SCEFR-AJT
    11. 12572
    Prev
    Next
    INDEX
    Explanations

    references to statistics and technical details, such as numbers, abbreviations, and specific data points

    oai_token-act-pair · gpt-3.5-turbo
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    neuronpedia/gpt2-small__res_scefr-ajt/6-res_scefr-ajt
    Prompts (Dashboard)
    12,288 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    46,080
    Data Type
    torch.float32
    Hook Point
    blocks.6.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    apollo-research/Skylion007-openwebtext-tokenizer-gpt2
    Hook Point Layer
    6
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    393
    +0.46
    2.0%
    385
    +0.18
    0.8%
    636
    +0.15
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    362
    +0.46
    0.26
    636
    +0.18
    0.21
    385
    +0.15
    0.25
    Negative Logits
     Judaism
    -0.77
     Catholicism
    -0.75
    bg
    -0.72
     Albania
    -0.71
    gat
    -0.70
    sted
    -0.69
    craft
    -0.68
    SPONSORED
    -0.68
    Lua
    -0.68
    /
    -0.68
    POSITIVE LOGITS
     same
    1.62
     entire
    1.59
     entirety
    1.48
    ses
    1.44
     slightest
    1.44
     latter
    1.40
     aforementioned
    1.39
     remainder
    1.39
     whole
    1.36
     hardest
    1.35
    Activations Density 4.363%

    No Known Activations