INDEX

Explanations

phrases and sentences discussing opinions, beliefs, and doubts

oai_token-act-pair · gpt-3.5-turbo

expressions of doubt or skepticism about opinions or statements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

itton

-0.68

pez

-0.67

asia

-0.66

ulner

-0.65

ERT

-0.63

 largeDownload

-0.57

 beware

-0.55

 whenever

-0.55

ahi

-0.55

hap

-0.55

POSITIVE LOGITS

 anymore

1.55

nor

1.16

 necessarily

1.12

 ever

1.11

any

1.09

 EVER

1.05

 slightest

1.03

 anything

1.00

 bothered

0.98

 anybody

0.93

Activations Density 0.222%

phrases and sentences discussing opinions, beliefs, and doubts

expressions of doubt or skepticism about opinions or statements

No Comments

No Known Activations