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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2026
    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 CE E2E Recon
    5. 6-RES_SCEFR-AJT
    6. 1097
    Prev
    Next
    INDEX
    Explanations

    words related to regulations and procedures, particularly in the context of financial agreements and legal matters

    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
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    253
    +0.13
    0.6%
    231
    +0.13
    0.6%
    746
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.13
    0.09
    231
    +0.13
    0.05
    326
    +0.12
    0.08
    Negative Logits
    ipple
    -0.70
    riched
    -0.65
    omething
    -0.65
    gets
    -0.61
    ktop
    -0.60
     mole
    -0.59
    mans
    -0.59
    cko
    -0.57
     misunder
    -0.56
    lights
    -0.56
    POSITIVE LOGITS
    rous
    0.84
    ctic
    0.82
    inity
    0.80
    riors
    0.79
    inal
    0.79
    ency
    0.78
    agin
    0.77
    pes
    0.77
    WAYS
    0.76
    ival
    0.75
    Activations Density 1.068%

    No Known Activations