© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    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. 1-RES-SM
    11. 28090
    Prev
    Next
    INDEX
    Explanations

    syntax structures and constructs in programming code

    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/1-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.1.hook_resid_post
    Hook Layer
    1
    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
    489
    +0.17
    1.0%
    434
    +0.13
    0.7%
    51
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    489
    +0.17
    0.04
    51
    +0.13
    0.02
    200
    +0.12
    0.03
    Negative Logits
    Ĩ
    -3.23
    ij
    -3.12
    ħ
    -3.10
    ĩ
    -3.03
    ĵ
    -3.03
    Ĺ
    -3.01
    ĻĤ
    -2.98
    £
    -2.97
    IJ
    -2.95
    ©
    -2.93
    POSITIVE LOGITS
     strangers
    1.65
     PUBLIC
    1.62
     couples
    1.59
     stranger
    1.58
    profit
    1.53
    er
    1.50
     prosecution
    1.43
     privilege
    1.43
     unexpected
    1.43
    erals
    1.43
    Activations Density 0.230%

    No Known Activations