© 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. 28323
    Prev
    Next
    INDEX
    Explanations

    references to technical specifications and programming elements

    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
    419
    +0.14
    0.8%
    274
    +0.12
    0.7%
    34
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    292
    +0.14
    0.16
    274
    +0.12
    0.19
    407
    +0.11
    0.11
    Negative Logits
    ĻĤ
    -2.17
    ¼
    -1.97
    Īĺ
    -1.58
    ŀ
    -1.56
    Ĵ
    -1.48
    ħ
    -1.45
    ĺ
    -1.44
    nez
    -1.44
    /+
    -1.43
    ľĵ
    -1.42
    POSITIVE LOGITS
    enix
    1.52
    amycin
    1.50
    angles
    1.47
    eton
    1.39
     ours
    1.35
    entre
    1.35
     fact
    1.31
    bracket
    1.29
    eking
    1.28
    amais
    1.28
    Activations Density 9.522%

    No Known Activations