INDEX

Explanations

instances of the word "glad" within the text

oai_token-act-pair · gpt-3.5-turbo

expressions of gratitude or contentment

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

classified

-0.78

$$$$

-0.71

Saharan

-0.67

improve

-0.66

artifacts

-0.65

heat

-0.64

irrel

-0.64

effic

-0.64

UE

-0.63

IDs

-0.62

POSITIVE LOGITS

 glad

1.30

 thankful

0.86

 happy

0.84

ness

0.81

 sorry

0.79

withstanding

0.79

itudinal

0.79

 grateful

0.79

 pardon

0.76

 thanking

0.74

Activations Density 0.006%

instances of the word "glad" within the text

expressions of gratitude or contentment

No Comments

No Known Activations