© 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. 2-RES-SM
    11. 10418
    Prev
    Next
    INDEX
    Explanations

    phrases that indicate possession or relational attributes

    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
    186
    +0.22
    1.2%
    23
    +0.13
    0.7%
    183
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    404
    +0.22
    0.08
    108
    +0.13
    0.09
    300
    +0.12
    0.07
    Negative Logits
     unfolded
    -1.90
    elike
    -1.69
    oyer
    -1.59
    feld
    -1.51
    thal
    -1.45
    iska
    -1.42
    ème
    -1.42
    æ°ı
    -1.41
     perturbed
    -1.39
    uary
    -1.39
    POSITIVE LOGITS
    ľĵ
    2.41
    ĻĤ
    2.25
    Īĺ
    1.67
    ĥ½
    1.62
    ÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤ
    1.51
    1.49
    ↵
    1.49
    ↵  âĢĥ
    1.49
    ↵
    1.49
    ↵
    1.49
    Activations Density 0.773%

    No Known Activations