INDEX

Explanations

phrases related to controversy or conflict

oai_token-act-pair · gpt-3.5-turbo

phrases indicating moral dilemmas or ethical conflicts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

ahu

-0.74

utral

-0.67

 Mechdragon

-0.66

iple

-0.63

USS

-0.61

Js

-0.58

é¾

-0.58

Wire

-0.57

¬¼

-0.57

 corrid

-0.57

POSITIVE LOGITS

 namely

1.22

 albeit

1.08

 despite

0.94

viz

0.93

nor

0.92

 especially

0.91

 irrespective

0.91

 regardless

0.90

 lest

0.90

especially

0.88

Activations Density 0.386%

phrases related to controversy or conflict

phrases indicating moral dilemmas or ethical conflicts

No Comments

No Known Activations