© 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. 6465
    Prev
    Next
    INDEX
    Explanations

    profanity and vulgar language

    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
    137
    +0.16
    0.6%
    48
    +0.14
    0.5%
    326
    +0.13
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    326
    +0.16
    0.06
    137
    +0.14
    0.04
    48
    +0.13
    0.04
    Negative Logits
    NetMessage
    -0.95
    HCR
    -0.90
    BIL
    -0.89
     Canaver
    -0.85
     conclud
    -0.84
    083
    -0.84
     reper
    -0.82
    eus
    -0.81
    anu
    -0.80
     Thib
    -0.79
    POSITIVE LOGITS
    holes
    1.23
     kidding
    1.21
    fuck
    1.16
     fuck
    1.13
     bastard
    1.02
     shit
    1.01
    amn
    1.01
     prick
    1.01
     fucking
    1.00
     FUCK
    0.99
    Activations Density 1.018%

    No Known Activations