INDEX

Explanations

mentions of names with the initials "J" or "J" followed by a number

oai_token-act-pair · gpt-3.5-turbo

instances of the character 'J' in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 offering

-0.65

arium

-0.63

fle

-0.63

ãĤ¼ãĤ¦ãĤ¹

-0.62

 containment

-0.61

 disc

-0.60

 buildup

-0.58

 tasting

-0.58

 uncovered

-0.57

 allot

-0.57

POSITIVE LOGITS

3.53

Js

2.36

JA

2.11

JR

2.04

JB

2.01

2.00

JD

1.99

1.96

JC

1.94

JV

1.91

Activations Density 0.017%

mentions of names with the initials "J" or "J" followed by a number

instances of the character 'J' in various contexts

No Comments

No Known Activations