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. 40267
    Prev
    Next
    INDEX
    Explanations

    names consisting of parts "hel" or "ley"

    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
    526
    +0.09
    0.5%
    293
    +0.07
    0.4%
    564
    +0.07
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    326
    +0.09
    0.19
    288
    +0.07
    0.17
    564
    +0.07
    0.07
    Negative Logits
    print
    -0.52
    ERC
    -0.48
     entangled
    -0.48
     infamous
    -0.47
    reme
    -0.47
    Rated
    -0.46
     Catalyst
    -0.45
     Sequ
    -0.45
     nomine
    -0.45
     newsp
    -0.45
    POSITIVE LOGITS
    tered
    0.85
    iflower
    0.78
    angelo
    0.73
    itably
    0.69
    ted
    0.69
    mand
    0.68
    brook
    0.67
    bos
    0.67
    tering
    0.67
    nikov
    0.66
    Activations Density 6.768%

    No Known Activations