© 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 Local
    9. 6-RES_SCL-AJT
    11. 13635
    Prev
    Next
    INDEX
    Explanations

    phrases related to adherence to beliefs or values

    oai_token-act-pair · gpt-3.5-turbo
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    neuronpedia/gpt2-small__res_scl-ajt/6-res_scl-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
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    549
    +0.06
    0.5%
    354
    +0.05
    0.4%
    293
    +0.05
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    393
    +0.06
    0.29
    640
    +0.05
    0.25
    375
    +0.05
    0.05
    Negative Logits
    ISON
    -0.51
    cheon
    -0.43
    ansas
    -0.41
    wat
    -0.41
    quished
    -0.40
    wcs
    -0.40
    mart
    -0.40
    ioch
    -0.40
     hoops
    -0.39
    allowed
    -0.38
    POSITIVE LOGITS
     Ashes
    0.44
     spite
    0.44
     Humanity
    0.43
     believer
    0.43
    onymous
    0.42
    oxide
    0.42
    isance
    0.40
    ggles
    0.39
     essence
    0.39
    DonaldTrump
    0.39
    Activations Density 12.310%

    No Known Activations