INDEX

Explanations

food-related words, particularly those related to pasta and noodles

oai_token-act-pair · gpt-3.5-turbo

references to pasta and noodle dishes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Tall

-0.77

 Hamp

-0.73

Ft

-0.72

izons

-0.71

 Austin

-0.71

sburg

-0.70

psy

-0.68

awks

-0.68

^^^^

-0.67

 Houston

-0.67

POSITIVE LOGITS

 noodles

1.22

 pasta

1.13

 nood

1.12

 spaghetti

0.97

arella

0.97

 slic

0.89

 soup

0.89

cake

0.85

 sauce

0.84

olini

0.83

Activations Density 0.008%

food-related words, particularly those related to pasta and noodles

references to pasta and noodle dishes

No Comments

No Known Activations