INDEX

Explanations

names of locations, specifically cities and countries

oai_token-act-pair · gpt-3.5-turbo

references to locations, particularly cities and geographical features

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

blance

-0.81

heit

-0.78

yip

-0.73

cliffe

-0.73

 Blumenthal

-0.70

 Eisenhower

-0.68

ufact

-0.67

letcher

-0.66

ittle

-0.64

 Blow

-0.64

POSITIVE LOGITS

 Janeiro

0.80

éŃĶ

0.70

ãĥ¼ãĥĨ

0.68

å£«

0.67

Ãĥ

0.64

 thous

0.64

ãĥŁ

0.63

Ïī

0.62

monary

0.61

ãĥ¼ãĥĨãĤ£

0.61

Activations Density 0.258%

names of locations, specifically cities and countries

references to locations, particularly cities and geographical features

No Comments

No Known Activations