INDEX

Explanations

terms related to politics and government

oai_token-act-pair · gpt-3.5-turbo

the character 'Ļ'

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

 gist

-0.65

 promoters

-0.64

anium

-0.62

 imitation

-0.61

odium

-0.57

 sophistication

-0.57

raviolet

-0.57

 range

-0.57

eanor

-0.57

 loopholes

-0.56

POSITIVE LOGITS

Ļ

1.29

å¥³

1.18

Ľ

1.02

ļ

1.00

ķ

0.99

ħ

0.99

çĲ

0.94

ı

0.94

º

0.92

İ

0.92

Activations Density 0.537%

terms related to politics and government

the character 'Ļ'

No Comments

No Known Activations