© 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 E2E Recon
    9. 6-RES_SCEFR-AJT
    11. 14672
    Prev
    Next
    INDEX
    Explanations

    statements emphasizing or asserting facts

    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
    192
    +0.15
    0.5%
    642
    +0.15
    0.5%
    718
    +0.13
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    192
    +0.15
    0.04
    544
    +0.15
    0.04
    697
    +0.13
    0.04
    Negative Logits
    lain
    -0.98
     Rumble
    -0.90
     lungs
    -0.86
    izont
    -0.85
    usable
    -0.84
    CLOSE
    -0.82
    hop
    -0.81
     Flavoring
    -0.80
     banana
    -0.77
    ather
    -0.76
    POSITIVE LOGITS
    ional
    1.34
    olkien
    1.01
     Eliot
    0.89
     Ezra
    0.89
    REP
    0.86
    herty
    0.85
    Versions
    0.84
    oln
    0.83
     Merrill
    0.78
    terday
    0.78
    Activations Density 0.504%

    No Known Activations