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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Apollo Research · Taylor · Identifying Functionally Important Features with End-to-End Sparse Dictionary Learning
    3. GPT2-Small
    4. Residual Similar CE Local
    5. 6-RES_SCL-AJT
    6. 13074
    Prev
    Next
    INDEX
    Explanations

    long character representations such as tokens and unique identifiers

    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
    447
    +0.14
    0.7%
    167
    +0.12
    0.6%
    255
    +0.08
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    167
    +0.14
    -0.03
    254
    +0.12
    0.05
    567
    +0.08
    0.05
    Negative Logits
    IED
    -0.59
    yg
    -0.57
    Trend
    -0.56
     Tant
    -0.56
    beam
    -0.56
    اÙĦ
    -0.55
    agles
    -0.55
     Norton
    -0.55
    ת
    -0.55
    ¿½
    -0.54
    POSITIVE LOGITS
    acters
    1.16
    istics
    1.01
    istically
    0.89
    istic
    0.64
    izations
    0.64
    inic
    0.60
    izes
    0.57
    alan
    0.57
    uria
    0.57
     sheet
    0.56
    Activations Density 6.949%

    No Known Activations