© 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
    2. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    3. Pythia-70M-Deduped
    4. Residuals
    5. 2-RES-SM
    6. 28385
    Prev
    Next
    INDEX
    Explanations

    dates, specifically focusing on historical events or anniversaries

    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
    286
    +0.13
    0.7%
    0
    +0.13
    0.7%
    178
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    286
    +0.13
    0.04
    0
    +0.13
    0.03
    21
    +0.13
    0.02
    Negative Logits
    ctors
    -1.70
    atics
    -1.58
    izens
    -1.50
    ights
    -1.48
    iquity
    -1.47
    ions
    -1.46
    ESULT
    -1.45
    rends
    -1.45
    osphere
    -1.43
    ressor
    -1.43
    POSITIVE LOGITS
    ĻĤ
    3.82
    ŀ
    2.78
    ĨĴ
    2.72
    IJ
    2.65
    Ŀ
    2.55
    Ļª
    2.46
    ĺ
    2.41
    ¬
    2.31
    ģ
    2.30
    º
    2.29
    Activations Density 0.161%

    No Known Activations