INDEX

Explanations

the word 'excuse' in various contexts

oai_token-act-pair · gpt-3.5-turbo

phrases that reference excuses or justifications

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

yss

-1.01

jri

-0.85

acid

-0.85

itals

-0.84

ovation

-0.81

hip

-0.80

duino

-0.78

ighth

-0.75

ickey

-0.72

hop

-0.72

POSITIVE LOGITS

 excuse

1.17

 excuses

1.16

 explanations

0.81

WHY

0.74

ãĤ¦ãĤ¹

0.74

ãĥ¼ãĥĨ

0.73

 justifying

0.73

 explanation

0.70

utical

0.70

 justification

0.70

Activations Density 0.018%

the word 'excuse' in various contexts

phrases that reference excuses or justifications

No Comments

No Known Activations