© 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
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 49152
    5. 8-RES_FS49152-JB
    6. 41772
    Prev
    Next
    INDEX
    Explanations

    pronouns and references indicating personal experiences or perceptions

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs49152-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    Data Type
    torch.float32
    Hook Point
    blocks.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    660
    +0.13
    0.6%
    671
    +0.10
    0.5%
    716
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    477
    +0.13
    0.02
    49
    +0.10
    0.03
    660
    +0.10
    0.02
    Negative Logits
     Demand
    -0.64
     Alas
    -0.61
     Bei
    -0.61
    US
    -0.61
    aly
    -0.60
    usterity
    -0.60
    æĪ¦
    -0.59
     Bots
    -0.58
    enium
    -0.58
    ardon
    -0.58
    POSITIVE LOGITS
     communicated
    0.96
     approach
    0.95
     communicate
    0.94
     handled
    0.88
     structured
    0.86
     communicates
    0.86
     handle
    0.82
     perceive
    0.82
     phr
    0.82
     handles
    0.81
    Activations Density 0.069%

    No Known Activations