INDEX

Explanations

references to governmental offices and political entities

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

phrases related to government institutions or positions

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

Arizona

-0.89

Minnesota

-0.88

Houston

-0.84

 Boise

-0.83

 Wichita

-0.81

Wisconsin

-0.80

Washington

-0.80

Seattle

-0.78

PHOTOS

-0.77

 Hawai

-0.77

POSITIVE LOGITS

 realise

1.10

 realised

1.07

 scept

1.02

abulary

1.02

 footballer

1.00

 apologised

1.00

 armour

1.00

 authorised

0.99

 analyse

0.97

 analys

0.96

Activations Density 0.204%

references to governmental offices and political entities

phrases related to government institutions or positions

No Comments

No Known Activations