INDEX

Explanations

Proper nouns related to awards or events

oai_token-act-pair · gpt-3.5-turbo

terms related to "priests" and religious titles

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

enegger

-0.93

icago

-0.83

schild

-0.78

lihood

-0.70

 bracelet

-0.70

 Anon

-0.70

 Beckham

-0.68

 HELL

-0.67

 Wink

-0.67

ç¥ŀ

-0.66

POSITIVE LOGITS

ests

1.07

Pri

1.05

zes

1.05

esses

1.02

Pri

1.01

ety

0.97

vy

0.88

eties

0.88

eur

0.87

quet

0.87

Activations Density 0.005%

Proper nouns related to awards or events

terms related to "priests" and religious titles

No Comments

No Known Activations