INDEX

Explanations

references to the city of Beijing

oai_token-act-pair · gpt-3.5-turbo

mentions of Beijing

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

oway

-0.73

adoes

-0.73

osate

-0.72

RH

-0.72

 Trooper

-0.69

pent

-0.68

âĢ¢âĢ¢âĢ¢âĢ¢

-0.68

ocene

-0.67

RANT

-0.67

 Hitchcock

-0.65

POSITIVE LOGITS

ijing

1.17

 Jinping

1.04

 Beijing

1.00

 Lumpur

0.99

 Yuan

0.89

jing

0.88

zhou

0.86

jin

0.84

 Jing

0.81

wei

0.81

Activations Density 0.012%

references to the city of Beijing

mentions of Beijing

No Comments

No Known Activations