© 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. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    5. Pythia-70M-Deduped
    7. Residuals
    9. 2-RES-SM
    11. 10947
    Prev
    Next
    INDEX
    Explanations

    occurrences of the word "with."

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    ctigges/pythia-70m-deduped__res-sm_processed/2-res-sm
    Prompts (Dashboard)
    32,768 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.2.hook_resid_post
    Hook Layer
    2
    Architecture
    standard
    Context Size
    128
    Dataset
    EleutherAI/the_pile_deduplicated
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    478
    +0.27
    1.5%
    358
    +0.14
    0.8%
    246
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    278
    +0.27
    0.05
    358
    +0.14
    0.06
    407
    +0.12
    0.05
    Negative Logits
    OUS
    -1.81
    thal
    -1.63
    blogger
    -1.53
     contributor
    -1.53
    ling
    -1.52
    stown
    -1.51
    ardo
    -1.50
    ship
    -1.49
    burg
    -1.49
    uler
    -1.47
    POSITIVE LOGITS
     respect
    3.17
     regard
    2.76
    ®
    2.49
    drawn
    2.34
    ¯
    2.33
     regards
    2.30
    Į
    2.03
    Ĭ
    2.00
    ĻĤ
    2.00
    ´
    1.96
    Activations Density 0.256%

    No Known Activations