INDEX

Explanations

keywords related to revealing or uncovering information

oai_token-act-pair · gpt-3.5-turbo

instances of the word "expose" and its variations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

wise

-0.73

erent

-0.67

rior

-0.65

reference

-0.65

tesy

-0.65

assian

-0.64

chief

-0.64

erva

-0.64

maker

-0.62

yip

-0.61

POSITIVE LOGITS

ibilities

0.86

 Breach

0.85

 weaknesses

0.83

Versions

0.82

é¾

0.79

 exposing

0.77

IBLE

0.77

Ð¼

0.72

ibility

0.72

 srfAttach

0.70

Activations Density 0.021%

keywords related to revealing or uncovering information

instances of the word "expose" and its variations

No Comments

No Known Activations