INDEX

Explanations

phrases indicating strong disapproval or rejection of something

oai_token-act-pair · gpt-3.5-turbo

terms related to criticism or condemnation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ammy

-0.80

omaly

-0.70

 captcha

-0.66

anuts

-0.66

olen

-0.66

OA

-0.65

 occupancy

-0.63

devices

-0.61

fficient

-0.60

OTA

-0.60

POSITIVE LOGITS

ounces

0.87

 denouncing

0.84

 disav

0.81

 denounce

0.80

 repud

0.75

 denounced

0.73

unci

0.73

uously

0.71

unciation

0.71

 Canaver

0.69

Activations Density 0.025%

phrases indicating strong disapproval or rejection of something

terms related to criticism or condemnation

No Comments

No Known Activations