INDEX

Explanations

phrases or words related to falsehoods, lies, or deception

oai_token-act-pair · gpt-3.5-turbo

the occurrence of the term "false" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

hens

-0.94

 guiActiveUnfocused

-0.81

hem

-0.78

xual

-0.77

rador

-0.73

shed

-0.72

onen

-0.71

icans

-0.71

okÃ©

-0.71

nesota

-0.70

POSITIVE LOGITS

 positives

1.24

 dich

1.03

 equival

0.98

 guiIcon

0.96

 accusation

0.95

 alarms

0.94

 negatives

0.91

 guiActiveUn

0.89

 allegation

0.85

 accusations

0.84

Activations Density 0.034%

phrases or words related to falsehoods, lies, or deception

the occurrence of the term "false" in various contexts

No Comments

No Known Activations