INDEX

Explanations

food-related words, especially vegetables

oai_token-act-pair · gpt-3.5-turbo

food items, particularly vegetables and potatoes

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ospace

-0.85

choes

-0.76

dinand

-0.73

instein

-0.72

lasses

-0.72

communications

-0.71

raints

-0.71

interrupted

-0.70

oustic

-0.69

POSITIVE LOGITS

 salads

1.24

 salad

1.24

 peel

1.21

 soup

1.09

 juice

1.07

 vegetables

1.03

 diced

1.03

 slices

1.02

 cabbage

1.01

 juices

1.01

Activations Density 0.049%

food-related words, especially vegetables

food items, particularly vegetables and potatoes

No Comments

No Known Activations