INDEX

Explanations

the word 'jud' followed by either a number or a word starting with a letter between 'a' and 'g'

oai_token-act-pair · gpt-3.5-turbo

occurrences of the term "jud" and its variations, indicating a focus on judgment-related concepts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

oteric

-0.72

Ø§ÙĦ

-0.64

xon

-0.64

Alz

-0.63

 Lyme

-0.63

ULAR

-0.63

 Niger

-0.62

 Afric

-0.61

CRE

-0.60

 Maurit

-0.60

POSITIVE LOGITS

gement

1.40

gments

1.25

Jud

1.16

gements

1.06

gment

1.06

jud

0.98

icator

0.98

iciary

0.96

itsu

0.95

jud

0.92

Activations Density 0.007%

the word 'jud' followed by either a number or a word starting with a letter between 'a' and 'g'

occurrences of the term "jud" and its variations, indicating a focus on judgment-related concepts

No Comments

No Known Activations