© 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. 4-RES-SM
    11. 12141
    Prev
    Next
    INDEX
    Explanations

    technical details related to software or digital content

    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/4-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.4.hook_resid_post
    Hook Layer
    4
    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
    271
    +0.14
    0.8%
    161
    +0.13
    0.7%
    423
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    255
    +0.14
    -0.01
    78
    +0.13
    -0.01
    271
    +0.11
    -0.00
    Negative Logits
    ing
    -1.52
    ter
    -1.44
    act
    -1.43
    enos
    -1.42
    Competing
    -1.39
    sen
    -1.39
    enberg
    -1.38
    enzyme
    -1.38
    enburg
    -1.37
    enstein
    -1.36
    POSITIVE LOGITS
    áĢ¯
    1.58
    ãģ¾ãģĽãĤĵ
    1.56
     true
    1.56
    ãģ¾ãģĽ
    1.52
    áĢ¬
    1.41
    à¹ī
    1.33
    fault
    1.30
    æģ¯
    1.29
    ÏĨο
    1.28
    оÑģÑģ
    1.24
    Activations Density 0.141%

    No Known Activations