INDEX

Explanations

contractions involving "wouldn't."

oai_token-act-pair · gpt-3.5-turbo

the phrase "wouldn't" in various contexts, indicating negation of an action or statement

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

 Gutenberg

-0.64

 Jelly

-0.63

 medi

-0.61

 Kills

-0.60

 spam

-0.60

PLA

-0.59

 nearest

-0.59

 princ

-0.58

 guided

-0.58

 couch

-0.57

POSITIVE LOGITS

't

1.45

uts

0.95

gered

0.94

itely

0.93

ÃŃ

0.92

eness

0.91

geon

0.87

emonium

0.86

ELY

0.86

ados

0.86

Activations Density 0.031%

contractions involving "wouldn't."

the phrase "wouldn't" in various contexts, indicating negation of an action or statement

No Comments

No Known Activations