© 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. 12498
    Prev
    Next
    INDEX
    Explanations

    structural elements and end tags in code or documentation

    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
    +0.16
    0.9%
    133
    +0.15
    0.8%
    225
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    133
    +0.16
    0.03
    0
    +0.15
    0.03
    173
    +0.11
    0.02
    Negative Logits
    ¨
    -2.59
    ¯
    -2.37
    Ł
    -2.36
    ĨĴ
    -2.21
    ĩ
    -2.07
    Ĥ¬
    -2.07
    ŀ
    -2.06
    ħ
    -2.05
    ĻĤ
    -2.04
    £
    -2.04
    POSITIVE LOGITS
    uvre
    1.59
     damned
    1.58
     insane
    1.55
     womb
    1.45
     bir
    1.41
     sav
    1.38
     impe
    1.34
     goddamn
    1.32
     endless
    1.32
     heartbeat
    1.32
    Activations Density 0.110%

    No Known Activations