© 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. 1-RES-SM
    11. 27252
    Prev
    Next
    INDEX
    Explanations

    activities related to research and examination processes

    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
    23
    +0.32
    1.9%
    156
    +0.19
    1.1%
    376
    +0.17
    1.0%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    198
    +0.32
    0.01
    148
    +0.19
    0.01
    45
    +0.17
    0.01
    Negative Logits
    heets
    -1.75
    arium
    -1.58
    ovic
    -1.55
     degli
    -1.54
     reductase
    -1.54
    hesis
    -1.51
    walks
    -1.51
    =”
    -1.49
     dei
    -1.47
     repertoire
    -1.47
    POSITIVE LOGITS
     how
    1.72
    them
    1.70
     authors
    1.70
    the
    1.63
    itely
    1.60
     the
    1.59
     them
    1.54
    these
    1.54
     temporal
    1.53
     what
    1.53
    Activations Density 0.067%

    No Known Activations