INDEX

Explanations

words related to theft or burglary

oai_token-act-pair · gpt-3.5-turbo

references to the city of Burgos

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

âĢ¢âĢ¢

-0.79

 Gemini

-0.72

 partName

-0.68

 filibuster

-0.63

 Apollo

-0.62

*/(

-0.61

hift

-0.61

 hypers

-0.58

 needle

-0.57

=-=-=-=-

-0.56

POSITIVE LOGITS

undy

1.25

 Burg

1.15

 burg

1.09

dor

0.95

rats

0.90

lar

0.90

erville

0.90

hers

0.79

ansk

0.79

nut

0.79

Activations Density 0.003%

words related to theft or burglary

references to the city of Burgos

No Comments

No Known Activations