© 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. 10797
    Prev
    Next
    INDEX
    Explanations

    instances of the word "miss" along with variations like "missed," "misses," and "missing."

    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
    491
    +0.18
    0.6%
    570
    +0.17
    0.6%
    651
    +0.15
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    491
    +0.18
    0.05
    562
    +0.17
    0.04
    518
    +0.15
    0.04
    Negative Logits
    advertisement
    -0.99
    Reviewer
    -0.97
     Jinping
    -0.97
    ayn
    -0.94
    loo
    -0.94
    Kar
    -0.86
    btn
    -0.85
    ipel
    -0.85
    pour
    -0.84
    gemony
    -0.84
    POSITIVE LOGITS
    pelled
    1.49
    pelling
    1.37
    hap
    1.35
    poke
    1.33
    ouri
    1.28
    pell
    1.25
    peak
    1.20
    pent
    1.00
    ibility
    0.99
    ordable
    0.98
    Activations Density 0.786%

    No Known Activations