Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    3. Pythia-70M-Deduped
    4. Residuals
    5. 1-RES-SM
    6. 26847
    Prev
    Next
    INDEX
    Explanations

    references to unit tests in coding

    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
    70
    +0.10
    0.6%
    200
    +0.10
    0.6%
    48
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    17
    +0.10
    0.01
    224
    +0.10
    0.03
    252
    +0.10
    0.02
    Negative Logits
    respectively
    -1.64
    itching
    -1.59
    uded
    -1.55
    instein
    -1.54
    inher
    -1.52
    soever
    -1.52
    uv
    -1.50
    ictor
    -1.49
    others
    -1.48
    ucing
    -1.48
    POSITIVE LOGITS
    heit
    2.01
    yards
    1.65
    yard
    1.54
    nement
    1.51
    esis
    1.50
    ário
    1.48
    ÅĽÄĩ
    1.47
    yrinth
    1.44
     Clause
    1.43
    anyahu
    1.42
    Activations Density 0.462%

    No Known Activations