© 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. OpenAI · Sparse Autoencoder for GPT2-Small - v5 2024
    5. GPT2-Small
    7. Residual Post - 32k
    9. 8-RES_POST_32K-OAI
    11. 4862
    Prev
    Next
    INDEX
    Explanations

    actions involving biting or taking a bite, particularly in a metaphorical or figurative sense

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_post_32k-oai
    Configuration
    jbloom/GPT2-Small-OAI-v5-32k-resid-post-SAEs/v5_32k_layer_8.pt
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.8.hook_resid_post
    Hook Layer
    8
    Architecture
    standard
    Context Size
    64
    Dataset
    Skylion007/openwebtext
    Activation Function
    topk
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    373
    +0.12
    0.6%
    154
    +0.12
    0.6%
    96
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    750
    +0.12
    0.04
    526
    +0.12
    0.00
    35
    +0.11
    0.03
    Negative Logits
    yrinth
    -0.76
    thumbnails
    -0.73
    soDeliveryDate
    -0.72
    HUD
    -0.71
    ENTION
    -0.69
    ATIONS
    -0.67
    amas
    -0.65
    TEXTURE
    -0.65
     prominently
    -0.63
    Sensor
    -0.62
    POSITIVE LOGITS
     flesh
    0.89
     bite
    0.87
     scraps
    0.86
     grapes
    0.86
     fing
    0.84
     bait
    0.83
     chunks
    0.83
     pulp
    0.83
     tongue
    0.81
     sucking
    0.80
    Activations Density 0.447%

    No Known Activations