© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI 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. 31367
    Prev
    Next
    INDEX
    Explanations

    references to legal issues and judicial proceedings

    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
    479
    +0.17
    1.0%
    150
    +0.12
    0.6%
    44
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    261
    +0.17
    0.06
    213
    +0.12
    0.07
    431
    +0.11
    0.06
    Negative Logits
    imedia
    -1.59
    ctrine
    -1.58
    "}](#
    -1.50
    uximab
    -1.48
    fficients
    -1.43
    saving
    -1.41
    ocardi
    -1.41
    izards
    -1.40
    isse
    -1.38
    vez
    -1.37
    POSITIVE LOGITS
     respectively
    1.79
    yr
    1.63
    £
    1.57
     age
    1.47
    ĨĴ
    1.46
     angry
    1.38
    anian
    1.38
    ļ
    1.38
    erated
    1.38
    ³
    1.35
    Activations Density 3.317%

    No Known Activations