© 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. 28250
    Prev
    Next
    INDEX
    Explanations

    actions and outcomes that have significant consequences or effects

    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
    186
    +0.17
    1.0%
    243
    +0.13
    0.8%
    47
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    243
    +0.17
    0.12
    323
    +0.13
    0.06
    510
    +0.13
    0.10
    Negative Logits
    "}](#
    -2.14
    ¿½
    -1.89
    ĻĤ
    -1.64
    ¹
    -1.51
     colleagues
    -1.50
    HS
    -1.47
    ¤
    -1.37
    «
    -1.37
    ±
    -1.37
    "."
    -1.37
    POSITIVE LOGITS
    ably
    1.84
    bbe
    1.75
    upon
    1.56
    blind
    1.54
    ificantly
    1.51
    weed
    1.50
    inally
    1.46
    ously
    1.45
    asan
    1.45
    yel
    1.44
    Activations Density 1.098%

    No Known Activations