INDEX

Explanations

Russian characters and symbols mixed with English words or phrases

oai_token-act-pair · gpt-3.5-turbo

the character "ļ" appearing multiple times

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

vironment

-0.85

cession

-0.80

 Beir

-0.79

estine

-0.79

sworth

-0.78

hran

-0.78

ijah

-0.76

 Hitman

-0.76

mercial

-0.75

onymous

-0.75

POSITIVE LOGITS

âķĲâķĲ

1.15

âķĲ

1.12

OUT

0.97

Ĺ

0.95

Ķ

0.87

pine

0.85

ÃĽ

0.83

à¤¾

0.83

RECT

0.83

cffffcc

0.81

Activations Density 0.013%

Russian characters and symbols mixed with English words or phrases

the character "ļ" appearing multiple times

No Comments

No Known Activations