INDEX

Explanations

phrases related to political discussions and positions

oai_token-act-pair · gpt-3.5-turbo

special characters or formatting elements in the text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 tremend

-0.75

 Moonlight

-0.65

aimon

-0.65

bable

-0.61

 conclud

-0.60

paio

-0.59

 Samar

-0.59

 Crusher

-0.58

 Burgess

-0.58

 Leopard

-0.57

POSITIVE LOGITS

ł

0.86

¹

0.76

ı

0.75

Į

0.70

į

0.70

Ķ

0.69

Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯

0.69

¼

0.67

º

0.67

¡

0.66

Activations Density 0.275%

phrases related to political discussions and positions

special characters or formatting elements in the text

No Comments

No Known Activations