© 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
    3. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    5. GPT2-Small
    7. Residual Stream
    9. 10-RES-JB
    11. 10545
    Prev
    Next
    INDEX
    Explanations

    evidence-related words and phrases such as "evidence," "statements," "DNA," and "forensic evidence."

    oai_token-act-pair · gpt-3.5-turbo

    references to evidence and investigative details in a narrative context

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 10-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    10
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    iversity
    -0.79
    Policy
    -0.79
    phabet
    -0.78
    aii
    -0.77
    liga
    -0.77
    utical
    -0.77
    utics
    -0.77
    Merit
    -0.76
    Congratulations
    -0.75
     ç¥ŀ
    -0.75
    POSITIVE LOGITS
     eyewitness
    1.46
     witnesses
    1.27
     gunshot
    1.24
     fingerprints
    1.22
     defendant
    1.14
     gunshots
    1.12
     autopsy
    1.12
     witness
    1.10
     gunman
    1.02
     forensic
    1.02
    Activations Density 0.428%

    No Known Activations