INDEX

Explanations

phrases related to a decrease or decline in something

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "down."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

ature

-0.80

Us

-0.76

ilies

-0.75

velt

-0.72

¶ħ

-0.71

Loading

-0.68

rehens

-0.68

xtap

-0.68

atures

-0.67

olkien

-0.67

POSITIVE LOGITS

 stairs

1.02

graded

0.98

stairs

0.86

grading

0.82

river

0.81

hill

0.80

played

0.79

 horm

0.78

cast

0.75

linked

0.74

Activations Density 0.022%

phrases related to a decrease or decline in something

occurrences of the word "down."

No Comments

No Known Activations