© 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. 10676
    Prev
    Next
    INDEX
    Explanations

    instances of the word "when" indicating temporal conditions or events

    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
    459
    +0.16
    0.9%
    376
    +0.11
    0.6%
    146
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    459
    +0.16
    0.08
    243
    +0.11
    0.07
    423
    +0.10
    0.06
    Negative Logits
    ITED
    -1.49
    et
    -1.42
    atory
    -1.41
    ored
    -1.39
     places
    -1.38
    ffected
    -1.38
    ray
    -1.37
    ksen
    -1.34
     Ont
    -1.32
    ior
    -1.32
    POSITIVE LOGITS
    ndef
    1.85
    rame
    1.60
     awake
    1.51
    llo
    1.49
    fy
    1.47
    soever
    1.46
     confronted
    1.43
     analys
    1.43
    ighs
    1.40
    blogspot
    1.38
    Activations Density 0.184%

    No Known Activations