INDEX

Explanations

phrases related to asserting or questioning the truth of statements

oai_token-act-pair · gpt-3.5-turbo

assertions of authenticity or truth regarding controversial subjects

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Cooldown

-0.80

ļéĨĴ

-0.77

Reviewer

-0.72

Marginal

-0.72

åĭ

-0.70

 reluct

-0.70

kward

-0.69

uyomi

-0.69

 margins

-0.69

oldown

-0.68

POSITIVE LOGITS

 hoax

1.23

 authentic

1.15

 genuine

1.13

 faked

1.12

 true

1.11

 legit

1.05

 real

1.04

 truth

1.04

 indeed

1.01

 fake

0.97

Activations Density 0.457%

phrases related to asserting or questioning the truth of statements

assertions of authenticity or truth regarding controversial subjects

No Comments

No Known Activations