© 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. 12326
    Prev
    Next
    INDEX
    Explanations

    components related to digital content organization and management

    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
    204
    +0.14
    0.8%
    218
    +0.12
    0.7%
    313
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    218
    +0.14
    0.10
    313
    +0.12
    0.06
    315
    +0.10
    0.04
    Negative Logits
    inally
    -1.74
    cyst
    -1.67
    inary
    -1.56
    mapsto
    -1.50
    ardless
    -1.50
    ulence
    -1.40
     theirs
    -1.38
    ging
    -1.30
    prote
    -1.28
    ){#
    -1.28
    POSITIVE LOGITS
    °
    3.51
    IJ
    3.36
    ¬
    3.35
    Ļ
    3.35
    ¥
    3.34
    İ
    3.34
    ¦
    3.34
    Ŀ
    3.28
    ³
    3.25
    ¾
    3.22
    Activations Density 3.043%

    No Known Activations