Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Apollo Research · Taylor · Identifying Functionally Important Features with End-to-End Sparse Dictionary Learning
    3. GPT2-Small
    4. Residual Similar L0 E2E Recon
    5. 2-RES_SLEFR-AJT
    6. 5397
    Prev
    Next
    INDEX
    Explanations

    mentions of notable individuals, particularly those with the last name Franklin or Booker

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    neuronpedia/gpt2-small__res_slefr-ajt/2-res_slefr-ajt
    Prompts (Dashboard)
    12,288 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    46,080
    Data Type
    torch.float32
    Hook Point
    blocks.2.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    apollo-research/Skylion007-openwebtext-tokenizer-gpt2
    Hook Point Layer
    2
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    247
    +0.08
    0.6%
    219
    +0.06
    0.5%
    649
    +0.06
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    189
    +0.08
    0.01
    430
    +0.06
    0.01
    377
    +0.06
    0.01
    Negative Logits
    ãĤº
    -0.51
    dust
    -0.44
    Advertisements
    -0.42
    clips
    -0.41
    ivation
    -0.39
    netflix
    -0.39
    cipline
    -0.38
    ARK
    -0.37
    -+-+-+-+
    -0.37
    ULAR
    -0.37
    POSITIVE LOGITS
     Franklin
    0.43
    geons
    0.39
    byss
    0.39
     encount
    0.38
     depress
    0.36
     Roosevelt
    0.36
     heav
    0.35
     Duff
    0.34
     Transit
    0.34
    ettings
    0.34
    Activations Density 0.028%

    No Known Activations