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. 29434
    Prev
    Next
    INDEX
    Explanations

    references to specific file paths or structures

    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
    156
    +0.26
    1.5%
    369
    +0.15
    0.9%
    466
    +0.15
    0.9%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    17
    +0.26
    0.00
    73
    +0.15
    0.01
    369
    +0.15
    0.01
    Negative Logits
    opters
    -1.61
    .]{}
    -1.56
    uro
    -1.52
    ONT
    -1.46
    romycin
    -1.46
     Algorithm
    -1.43
    udson
    -1.41
    ousseau
    -1.39
    ruitment
    -1.36
    PLIED
    -1.35
    POSITIVE LOGITS
     soda
    1.67
    rell
    1.58
    icas
    1.56
    hammer
    1.56
    dom
    1.55
    bar
    1.54
    leg
    1.53
     cigar
    1.47
    bell
    1.45
    fight
    1.44
    Activations Density 0.114%

    No Known Activations