INDEX

Explanations

phrases indicating a negative context or opposition

oai_token-act-pair · gpt-3.5-turbo

the word "Not" and its variations in different contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

velt

-0.70

stakes

-0.67

hower

-0.61

ç·

-0.60

æĸ

-0.59

 Circuit

-0.59

kamp

-0.58

CHR

-0.57

Kry

-0.56

ixel

-0.56

POSITIVE LOGITS

epad

1.33

eworthy

1.16

icably

1.12

withstanding

1.11

icable

1.10

 necessarily

1.08

orious

1.08

ifications

1.06

ifier

1.04

ional

1.01

Activations Density 0.102%

phrases indicating a negative context or opposition

the word "Not" and its variations in different contexts

No Comments

No Known Activations