INDEX

Explanations

specific non-English characters in a document, particularly those representing the Japanese Yen symbol

oai_token-act-pair · gpt-3.5-turbo

occurrences of the symbol "¥" and associated graphical artifacts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 circulation

-0.83

 suspensions

-0.79

 primates

-0.79

hyde

-0.76

 subpoen

-0.74

 consolidation

-0.73

 distracted

-0.73

 mammals

-0.72

 overlapping

-0.72

 deepening

-0.71

POSITIVE LOGITS

ï¸

1.29

ÙĨ

1.12

Xi

1.09

¥

1.07

ÙĲ

1.07

ÙĦ

1.04

×ķ

1.04

×ľ

1.03

Ùħ

1.01

à¤¾

0.99

Activations Density 0.007%

specific non-English characters in a document, particularly those representing the Japanese Yen symbol

occurrences of the symbol "¥" and associated graphical artifacts

No Comments

No Known Activations