© 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. 5056
    Prev
    Next
    INDEX
    Explanations

    steps or instructions to complete a specific task or achieve a result

    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
    447
    +0.24
    1.0%
    393
    +0.13
    0.5%
    78
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    640
    +0.24
    0.30
    521
    +0.13
    0.04
    64
    +0.11
    -0.22
    Negative Logits
    inventoryQuantity
    -0.86
     Flavoring
    -0.80
     Dur
    -0.75
     Tru
    -0.71
     Claims
    -0.69
    Magikarp
    -0.65
    Books
    -0.65
     Represent
    -0.65
    Chance
    -0.65
     Vehicles
    -0.64
    POSITIVE LOGITS
     abundantly
    0.84
     job
    0.83
     anecd
    0.82
     unconsciously
    0.81
     safely
    0.79
     efficiently
    0.79
     entails
    0.79
    berra
    0.79
     toget
    0.79
     cheaply
    0.78
    Activations Density 10.954%

    No Known Activations