© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    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 Local
    9. 6-RES_SCL-AJT
    11. 13953
    Prev
    Next
    INDEX
    Explanations

    words related to a television show or series

    oai_token-act-pair · gpt-3.5-turbo
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    neuronpedia/gpt2-small__res_scl-ajt/6-res_scl-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
    739
    +0.08
    0.5%
    679
    +0.07
    0.5%
    86
    +0.07
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    433
    +0.08
    0.09
    247
    +0.07
    0.07
    151
    +0.07
    0.15
    Negative Logits
     punishing
    -0.50
    xon
    -0.50
    ãĥ¢
    -0.49
     Mehran
    -0.47
    女
    -0.46
    mit
    -0.46
     sled
    -0.45
    itan
    -0.45
     vengeance
    -0.44
     seams
    -0.44
    POSITIVE LOGITS
    edIn
    0.83
    pedia
    0.62
    eous
    0.56
     Crimes
    0.55
     Groups
    0.54
    WI
    0.54
     Vector
    0.52
    ancies
    0.51
     Heights
    0.51
     Interest
    0.51
    Activations Density 7.997%

    No Known Activations