INDEX

Explanations

numerical patterns within text

oai_token-act-pair · gpt-3.5-turbo

the occurrence of numerical figures or statistics in the text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 travers

-0.70

 swinging

-0.68

hung

-0.68

 heroine

-0.66

 guarding

-0.66

 brav

-0.65

 rebuilding

-0.64

 hust

-0.63

 swallowing

-0.63

 railways

-0.62

POSITIVE LOGITS

 SHARES

0.97

ctr

0.94

maxwell

0.89

:{

0.82

 Expand

0.81

å¹

0.81

 Explicit

0.80

 ILCS

0.79

Reviewer

0.78

MN

0.76

Activations Density 0.128%

numerical patterns within text

the occurrence of numerical figures or statistics in the text

No Comments

No Known Activations