INDEX

Explanations

disclaimers at the beginning or end of texts

oai_token-act-pair · gpt-3.5-turbo

phrases that emphasize disclaimers and the non-endorsement of views or products

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

footed

-0.80

kens

-0.75

 yawn

-0.71

window

-0.71

saf

-0.70

noticed

-0.68

iddler

-0.67

 srfAttach

-0.67

éŃĶ

-0.62

 reperto

-0.61

POSITIVE LOGITS

 condone

1.25

 endorse

1.14

 affiliated

1.12

 endorsement

1.10

 imply

0.99

 endorsements

0.96

 endorsing

0.93

nor

0.89

 spam

0.89

 endorsed

0.88

Activations Density 0.241%

disclaimers at the beginning or end of texts

phrases that emphasize disclaimers and the non-endorsement of views or products

No Comments

No Known Activations