Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Apollo Research · Taylor · Identifying Functionally Important Features with End-to-End Sparse Dictionary Learning
    3. GPT2-Small
    4. Residual Similar CE Local
    5. 6-RES_SCL-AJT
    6. 13411
    Prev
    Next
    INDEX
    Explanations

    quotes and phrases within quotation marks

    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
    607
    +0.03
    0.5%
    7
    +0.03
    0.5%
    236
    +0.03
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    607
    +0.03
    0.09
    236
    +0.03
    0.11
    108
    +0.03
    0.07
    Negative Logits
    ̶
    -0.20
    ounty
    -0.18
    Editor
    -0.18
    amins
    -0.18
    Translation
    -0.17
    nered
    -0.17
     Dragonbound
    -0.17
    mares
    -0.17
    ãĤ¿
    -0.17
     Izan
    -0.16
    POSITIVE LOGITS
    iband
    0.22
    icit
    0.22
    miah
    0.21
    ittle
    0.18
    itary
    0.18
    mosp
    0.18
    ````
    0.18
    apple
    0.18
     WATCHED
    0.17
     `
    0.17
    Activations Density 9.149%

    No Known Activations