INDEX

Explanations

phrases related to man-made issues or conflicts

oai_token-act-pair · gpt-3.5-turbo

references to male-related vulnerabilities and conflicts

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

BIL

-0.75

nces

-0.73

Birth

-0.70

DER

-0.70

aminer

-0.69

DragonMagazine

-0.69

Impl

-0.68

Requ

-0.67

ascript

-0.67

taboola

-0.67

POSITIVE LOGITS

ape

0.71

 wiser

0.70

ghai

0.70

bane

0.67

 Scotch

0.66

 apes

0.66

 Viet

0.66

coni

0.65

 Samoa

0.65

 circumcised

0.64

Activations Density 0.485%

phrases related to man-made issues or conflicts

references to male-related vulnerabilities and conflicts

No Comments

No Known Activations