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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    2. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    3. Pythia-70M-Deduped
    4. Residuals
    5. 2-RES-SM
    6. 28683
    Prev
    Next
    INDEX
    Explanations

    references to hosts and their relationships with parasites

    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
    156
    +0.17
    1.0%
    351
    +0.12
    0.7%
    127
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    351
    +0.17
    0.02
    481
    +0.12
    0.02
    277
    +0.10
    0.02
    Negative Logits
     sue
    -1.64
     remand
    -1.58
     progress
    -1.45
     fprintf
    -1.44
     hope
    -1.43
     oligonucle
    -1.42
     mean
    -1.42
     certain
    -1.41
    ){#
    -1.37
     suture
    -1.37
    POSITIVE LOGITS
    ilities
    3.09
    esses
    2.22
    eles
    2.11
    ility
    2.10
    yard
    2.03
    ages
    1.91
    grounds
    1.91
    etter
    1.81
    ess
    1.80
    LLY
    1.76
    Activations Density 0.111%

    No Known Activations