INDEX

Explanations

dates, years, and periods of time

oai_token-act-pair · gpt-3.5-turbo

occurrences of numerical information and time references within the text

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Widget

-0.78

Quantity

-0.78

ãĤ¨ãĥ«

-0.77

ļéĨĴ

-0.76

zbollah

-0.76

Otherwise

-0.75

Sov

-0.74

Higher

-0.73

tty

-0.71

FIX

-0.71

POSITIVE LOGITS

he

1.59

she

1.11

his

1.08

 however

1.00

 they

0.95

he

0.92

him

0.91

 including

0.81

 though

0.79

 whenever

0.79

Activations Density 0.246%

dates, years, and periods of time

occurrences of numerical information and time references within the text

No Comments

No Known Activations