© 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. 12284
    Prev
    Next
    INDEX
    Explanations

    expressions of agreement or consensus

    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
    365
    +0.13
    0.7%
    29
    +0.12
    0.7%
    329
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    329
    +0.13
    0.02
    367
    +0.12
    0.02
    29
    +0.11
    0.01
    Negative Logits
     imagination
    -1.77
    elves
    -1.59
     tomorrow
    -1.55
     talents
    -1.55
    uka
    -1.45
    »
    -1.43
    icum
    -1.41
    omycin
    -1.40
    ¹
    -1.39
     Studies
    -1.39
    POSITIVE LOGITS
    ably
    2.15
    ableness
    1.84
    riterion
    1.69
    atively
    1.53
     arbit
    1.52
    arser
    1.51
    able
    1.44
    keepers
    1.42
    lement
    1.42
     stipulation
    1.42
    Activations Density 0.211%

    No Known Activations