INDEX

Explanations

news-related entities like names of individuals, places, and organizations

oai_token-act-pair · gpt-3.5-turbo

legal terms and phrases associated with crime and legal proceedings

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

ĸļ

-0.75

uph

-0.72

©¶æ

-0.67

 weekends

-0.66

 purs

-0.65

 everyday

-0.64

 covenant

-0.64

 ourselves

-0.64

 ability

-0.63

¬¼

-0.63

POSITIVE LOGITS

ccording

1.15

According

1.04

Writing

1.03

Asked

1.02

Among

1.02

SPONSORED

1.00

However

0.99

Sources

0.98

But

0.98

Their

0.98

Activations Density 0.542%

news-related entities like names of individuals, places, and organizations

legal terms and phrases associated with crime and legal proceedings

No Comments

No Known Activations