INDEX

Explanations

proper names

oai_token-act-pair · gpt-3.5-turbo

proper nouns and names

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

illance

-0.70

uyomi

-0.69

upuncture

-0.66

izoph

-0.64

Interstitial

-0.64

direction

-0.63

¥ŀ

-0.62

opsy

-0.61

 dystopian

-0.61

 surv

-0.61

POSITIVE LOGITS

 Larson

1.09

 McKay

1.04

 Coleman

1.03

 Roberts

1.01

 Robinson

1.01

 Hoffman

1.00

 McCull

1.00

McC

0.99

 Cumm

0.96

 Richards

0.96

Activations Density 0.242%

proper names

proper nouns and names

No Comments

No Known Activations