© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    5. Pythia-70M-Deduped
    7. Residuals
    9. 1-RES-SM
    11. 28917
    Prev
    Next
    INDEX
    Explanations

    past-tense verbs and their variations in different forms

    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/1-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.1.hook_resid_post
    Hook Layer
    1
    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
    156
    +0.19
    1.1%
    488
    +0.14
    0.8%
    493
    +0.14
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    493
    +0.19
    0.08
    488
    +0.14
    0.09
    22
    +0.14
    0.06
    Negative Logits
    lessly
    -1.82
     varied
    -1.39
     smelled
    -1.38
    aming
    -1.36
     dawn
    -1.35
    ..."
    -1.34
    cases
    -1.32
     intimid
    -1.31
     %
    -1.29
     \"
    -1.29
    POSITIVE LOGITS
    irectory
    2.08
    ĻĤ
    2.06
    icense
    1.92
    pez
    1.91
    don
    1.84
    ialog
    1.79
    ocs
    1.74
    icine
    1.74
    ¸
    1.71
    entary
    1.68
    Activations Density 0.438%

    No Known Activations