© 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. 30118
    Prev
    Next
    INDEX
    Explanations

    expressions of asking for information or feedback

    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
    +0.27
    1.5%
    320
    +0.17
    1.0%
    217
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    419
    +0.27
    0.04
    320
    +0.17
    0.04
    257
    +0.11
    0.03
    Negative Logits
    ĻĤ
    -1.93
    Ļª
    -1.70
    ĨĴ
    -1.46
    ŀ
    -1.44
    ļ
    -1.43
     features
    -1.37
    elijk
    -1.36
    chap
    -1.35
    STEM
    -1.35
     imagin
    -1.34
    POSITIVE LOGITS
    inate
    1.69
     disclose
    1.56
    omycin
    1.52
     cond
    1.50
    quez
    1.49
    uce
    1.48
    quote
    1.45
     judge
    1.42
     appropriately
    1.42
    opolys
    1.41
    Activations Density 0.287%

    No Known Activations