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. 2-RES-SM
    6. 11799
    Prev
    Next
    INDEX
    Explanations

    references to public service and government-related activities

    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
    419
    +0.19
    1.1%
    330
    +0.17
    1.0%
    57
    +0.14
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    111
    +0.19
    0.21
    71
    +0.17
    0.23
    156
    +0.14
    0.23
    Negative Logits
    ---|---
    -1.81
    ---|
    -1.69
    ---|---|---
    -1.57
    rained
    -1.56
    hest
    -1.55
    oler
    -1.49
    foreach
    -1.46
    iles
    -1.43
    '"
    -1.41
    most
    -1.38
    POSITIVE LOGITS
    etary
    1.46
     Caption
    1.43
     Markets
    1.43
     avenues
    1.38
    esis
    1.36
     Bulletin
    1.33
     caption
    1.33
     opinions
    1.32
     Draft
    1.32
     comment
    1.32
    Activations Density 3.873%

    No Known Activations