INDEX

Explanations

phrases related to warnings or being cautious

oai_token-act-pair · gpt-3.5-turbo

phrases emphasizing the concept of caution or carefulness

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

upon

-0.78

amen

-0.77

ono

-0.73

ibur

-0.70

inished

-0.69

phabet

-0.64

 apologized

-0.63

 Canad

-0.63

flat

-0.62

 apology

-0.62

POSITIVE LOGITS

 lest

1.05

 pitfalls

0.71

winter

0.71

 calibr

0.69

Yen

0.69

rored

0.66

ricanes

0.65

idious

0.62

ergic

0.61

esson

0.61

Activations Density 0.134%

phrases related to warnings or being cautious

phrases emphasizing the concept of caution or carefulness

No Comments

No Known Activations