INDEX

Explanations

references to the country "France."

oai_token-act-pair · gpt-3.5-turbo

mentions of France

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

iary

-0.92

regor

-0.88

atari

-0.81

iating

-0.80

âĹ¼

-0.78

ividual

-0.78

uilt

-0.77

ramid

-0.77

iated

-0.77

iations

-0.77

POSITIVE LOGITS

 Alps

0.87

 Hollande

0.83

 fries

0.82

 France

0.80

Riv

0.79

 MÃ©

0.76

 Marse

0.75

France

0.72

 Franc

0.72

 countryside

0.69

Activations Density 0.016%

references to the country "France."

mentions of France

No Comments

No Known Activations