INDEX

Explanations

questions or statements expressing curiosity or seeking information

oai_token-act-pair · gpt-3.5-turbo

inquiries and questions about knowledge or information

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

Negative Logits

©¶æ¥µ

-0.69

Introduced

-0.66

twitch

-0.62

edition

-0.61

ufact

-0.60

absor

-0.60

jam

-0.60

oun

-0.58

Pear

-0.58

 understatement

-0.57

POSITIVE LOGITS

 whether

1.32

why

1.27

how

1.25

WHY

1.23

whether

1.06

 specifics

1.05

 answers

1.01

why

1.00

 what

1.00

HOW

0.98

Activations Density 0.123%

questions or statements expressing curiosity or seeking information

inquiries and questions about knowledge or information

No Known Activations