© 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. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    5. GPT2-Small
    7. Residual Stream
    9. 2-RES-JB
    11. 4269
    Prev
    Next
    INDEX
    Explanations

    words related to eccentricity

    oai_token-act-pair · gpt-3.5-turbo

    descriptors related to unconventional or unusual characteristics

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 2-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.hook_resid_pre
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    Data Type
    torch.float32
    Hook Point
    blocks.2.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    2
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    utenant
    -0.73
    uden
    -0.70
     abund
    -0.66
     saliva
    -0.64
     Oxy
    -0.62
    onds
    -0.60
    rils
    -0.60
    elsen
    -0.60
     Nare
    -0.60
     guid
    -0.60
    POSITIVE LOGITS
    entric
    0.86
    ity
    0.85
    rift
    0.82
     eccentric
    0.82
    ities
    0.80
    isc
    0.75
    ishly
    0.73
    Stretch
    0.72
    wra
    0.72
    verted
    0.72
    Activations Density 0.025%

    No Known Activations