© 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. 30323
    Prev
    Next
    INDEX
    Explanations

    instances of brackets or list-like 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
    478
    +0.23
    1.3%
    103
    +0.12
    0.7%
    51
    +0.11
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    103
    +0.23
    0.04
    51
    +0.12
    0.03
    128
    +0.11
    0.02
    Negative Logits
    iture
    -1.82
    itures
    -1.63
    uates
    -1.54
    sit
    -1.51
    ading
    -1.50
    xture
    -1.47
    pins
    -1.46
     attack
    -1.45
     England
    -1.44
     surrounding
    -1.43
    POSITIVE LOGITS
    Ĵ
    2.79
    ī
    2.63
    İ
    2.60
    IJ
    2.52
    Ĭ
    2.47
    Ķ
    2.45
    Ģ
    2.34
    unreadable
    2.32
    ŀ
    2.07
    Į
    2.06
    Activations Density 0.079%

    No Known Activations