INDEX

Explanations

phrases related to commands or instructions, often with negative connotations

oai_token-act-pair · gpt-3.5-turbo

the phrase "don't" 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 @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

çīĪ

-0.80

 Palest

-0.79

anwhile

-0.71

RAD

-0.70

 Blaz

-0.65

è»

-0.63

 Leilan

-0.63

BuyableInstoreAndOnline

-0.61

 Accessed

-0.61

 Trojan

-0.60

POSITIVE LOGITS

£

1.06

¿

1.02

Ķ

0.98

ĸ

0.98

¬

0.96

Ĭ

0.96

į

0.94

¼

0.93

ĵ

0.93

¡

0.93

Activations Density 0.171%

phrases related to commands or instructions, often with negative connotations

the phrase "don't" in various contexts

No Comments

No Known Activations