© 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 EvalsExportsinterp-engineNEWAPI 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. 2359
    Prev
    Next
    INDEX
    Explanations

    statements or descriptions involving the concept of "without" followed by a noun or verb

    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
    616
    +0.17
    0.5%
    238
    +0.16
    0.5%
    455
    +0.14
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    238
    +0.17
    0.06
    669
    +0.16
    0.05
    471
    +0.14
    0.05
    Negative Logits
    geist
    -1.24
    DragonMagazine
    -1.07
    oola
    -1.05
    lish
    -1.02
    omsky
    -0.99
    =-=-=-=-=-=-=-=-
    -0.99
     murd
    -0.98
    mers
    -0.97
    aven
    -0.96
     Coy
    -0.96
    POSITIVE LOGITS
     regard
    1.37
     mentioning
    1.17
    lessly
    1.10
     compromising
    1.10
    Charges
    1.09
     necessarily
    1.07
     hesitation
    1.04
     touching
    1.03
     exception
    1.03
     adequate
    1.03
    Activations Density 0.672%

    No Known Activations