© 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. 2-RES-SM
    11. 28505
    Prev
    Next
    INDEX
    Explanations

    references to financial support and funding 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
    156
    +0.11
    0.6%
    174
    +0.11
    0.6%
    186
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    186
    +0.11
    0.02
    141
    +0.11
    0.03
    7
    +0.11
    0.01
    Negative Logits
     pleas
    -1.84
     queer
    -1.77
    Ļª
    -1.76
     yell
    -1.72
    ĨĴ
    -1.70
    atin
    -1.68
     familiar
    -1.63
     dialect
    -1.62
    plify
    -1.59
     rude
    -1.59
    POSITIVE LOGITS
    raiser
    2.49
    amental
    2.14
    raised
    2.10
    rais
    2.08
    actor
    2.01
    raising
    1.96
    orship
    1.81
    raise
    1.79
     allocation
    1.79
    ament
    1.76
    Activations Density 0.293%

    No Known Activations