INDEX

Explanations

numerical figures

oai_token-act-pair · gpt-3.5-turbo

numerical values and statistics in the text

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 eleph

-1.15

Þ

-1.12

 exting

-1.06

 pione

-1.03

ortunately

-1.03

aditional

-1.02

ß

-0.98

ò

-0.97

ccording

-0.96

oun

-0.95

POSITIVE LOGITS

igg

0.52

athan

0.50

inx

0.49

ratch

0.48

âĶĢâĶĢâĶĢâĶĢ

0.48

outheastern

0.48

Course

0.48

't

0.46

Ð°

0.46

istar

0.46

Activations Density 1.628%

numerical figures

numerical values and statistics in the text

No Comments

No Known Activations