© 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. 11193
    Prev
    Next
    INDEX
    Explanations

    specific identifiers and elements related to coding or data structures.

    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
    436
    +0.12
    0.7%
    335
    +0.11
    0.6%
    423
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    111
    +0.12
    0.39
    156
    +0.11
    0.39
    71
    +0.10
    0.38
    Negative Logits
     there
    -1.49
     enough
    -1.49
     quite
    -1.40
     depths
    -1.36
    ows
    -1.33
    !),
    -1.30
     resistivity
    -1.30
    ingly
    -1.28
     even
    -1.28
     --
    -1.27
    POSITIVE LOGITS
    anine
    1.59
    __(
    1.59
    bracket
    1.45
     quo
    1.45
    ophyll
    1.42
    reon
    1.41
    bsite
    1.41
    fluence
    1.37
    veolar
    1.37
    Xi
    1.36
    Activations Density 7.194%

    No Known Activations