INDEX

Explanations

words and prefixes related to medical conditions

oai_token-act-pair · gpt-3.5-turbo

the presence of the substring "ec" within words

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

 dear

-0.66

inson

-0.59

 hitters

-0.58

åĳ

-0.58

 Rack

-0.56

rug

-0.56

>>>>

-0.56

Ily

-0.56

 Desire

-0.54

 Rouse

-0.53

POSITIVE LOGITS

stasy

1.24

uador

1.21

lectic

1.17

essors

1.05

utive

1.04

chio

1.01

ourse

0.98

entric

0.97

heon

0.96

essor

0.96

Activations Density 0.017%

words and prefixes related to medical conditions

the presence of the substring "ec" within words

No Comments

No Known Activations