© 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. 13084
    Prev
    Next
    INDEX
    Explanations

    references to financial markets and institutions

    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
    459
    +0.14
    0.7%
    642
    +0.13
    0.6%
    311
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    311
    +0.14
    -0.04
    326
    +0.13
    0.21
    145
    +0.10
    0.07
    Negative Logits
    76561
    -0.84
    resent
    -0.77
    cific
    -0.75
    CVE
    -0.74
    cffffcc
    -0.70
    senal
    -0.66
    mur
    -0.65
    hig
    -0.64
    cig
    -0.63
    iard
    -0.62
    POSITIVE LOGITS
     Catalog
    0.81
     deviation
    0.70
     Format
    0.68
    ity
    0.67
     Fargo
    0.66
    ature
    0.66
     Template
    0.65
    orrow
    0.65
     Definition
    0.65
     Breakfast
    0.65
    Activations Density 7.998%

    No Known Activations