© 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
    3. Apollo Research · Taylor · Identifying Functionally Important Features with End-to-End Sparse Dictionary Learning
    5. GPT2-Small
    7. Residual Similar CE E2E Recon
    9. 6-RES_SCEFR-AJT
    11. 7
    Prev
    Next
    INDEX
    Explanations

    references to ownership or being an owner

    oai_token-act-pair · gpt-3.5-turbo
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    neuronpedia/gpt2-small__res_scefr-ajt/6-res_scefr-ajt
    Prompts (Dashboard)
    12,288 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    46,080
    Data Type
    torch.float32
    Hook Point
    blocks.6.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    apollo-research/Skylion007-openwebtext-tokenizer-gpt2
    Hook Point Layer
    6
    Activation Function
    relu
    Embeds
    IFrame
    Link
    •test_api
    You can add features from any model or SAE in one list.
    •test_api23
    You can add features from any model or SAE in one list.
    •test_api5
    You can add features from any model or SAE in one list.
    •test_api6
    You can add features from any model or SAE in one list.
    •test_api7
    You can add features from any model or SAE in one list.

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    167
    +0.17
    0.6%
    535
    +0.16
    0.6%
    224
    +0.14
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    224
    +0.17
    0.06
    146
    +0.16
    0.05
    32
    +0.14
    0.05
    Negative Logits
     Airl
    -1.18
     RELEASE
    -1.01
    eday
    -0.98
     Nadu
    -0.96
    bil
    -0.94
    enegger
    -0.92
    inx
    -0.88
    utton
    -0.87
    ilde
    -0.87
    schild
    -0.86
    POSITIVE LOGITS
    pers
    1.05
    hip
    1.01
    owners
    1.01
     beware
    0.99
    ership
    0.94
     owner
    0.93
     license
    0.93
    owner
    0.90
    lege
    0.90
    fred
    0.90
    Activations Density 0.749%

    No Known Activations