INDEX

Explanations

phrases or sentences containing the word "misleading."

oai_token-act-pair · gpt-3.5-turbo

terms associated with misinformation and misleading information

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

dain

-0.70

DAY

-0.67

mun

-0.67

itar

-0.66

empl

-0.65

inct

-0.64

El

-0.63

%"

-0.63

Goal

-0.62

examination

-0.62

POSITIVE LOGITS

 misled

1.33

 mislead

1.31

 misleading

1.25

 deceive

1.14

 deceived

1.05

 dece

0.99

ingly

0.95

 confuse

0.95

 misrepresent

0.90

 falsely

0.84

Activations Density 0.018%

phrases or sentences containing the word "misleading."

terms associated with misinformation and misleading information

No Comments

No Known Activations