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 E2E Recon
    5. 6-RES_SCEFR-AJT
    6. 534
    Prev
    Next
    INDEX
    Explanations

    verbs in the infinitive form occurring after prepositions

    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
    718
    +0.20
    0.7%
    393
    +0.20
    0.7%
    608
    +0.18
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    718
    +0.20
    0.11
    608
    +0.20
    0.11
    665
    +0.18
    0.10
    Negative Logits
     Used
    -0.94
    Used
    -0.88
     Ended
    -0.87
     Drops
    -0.86
    prem
    -0.85
     Entered
    -0.82
     Uses
    -0.81
     Accessed
    -0.81
     Flo
    -0.79
     Indust
    -0.76
    POSITIVE LOGITS
     ensure
    1.38
     promote
    1.33
     improve
    1.32
     prevent
    1.31
     emulate
    1.31
     avoid
    1.28
    pless
    1.23
     alleviate
    1.23
     reduce
    1.23
     demonstrate
    1.22
    Activations Density 0.976%

    No Known Activations