INDEX

Explanations

words related to reviews and opinions

oai_token-act-pair · gpt-3.5-turbo

references to specific individuals or commentators providing opinions or analyses

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 postponed

-0.71

oyer

-0.66

ipal

-0.66

uez

-0.65

\'

-0.64

iances

-0.63

 Pradesh

-0.63

addons

-0.62

ARA

-0.60

 customs

-0.60

POSITIVE LOGITS

 eloqu

1.28

 insightful

1.18

 fascinating

1.15

 brilliantly

1.13

 succinct

1.11

 debunk

1.07

 excellent

1.06

 terrific

1.04

 hilarious

1.03

 aptly

1.01

Activations Density 0.427%

words related to reviews and opinions

references to specific individuals or commentators providing opinions or analyses

No Comments

No Known Activations