INDEX

Explanations

information related to Poland

oai_token-act-pair · gpt-3.5-turbo

references to geographic locations and their significance

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Austral

-1.03

Haw

-0.98

 SpaceX

-0.98

 Hawth

-0.97

Houston

-0.97

LSU

-0.96

 Falcon

-0.94

heed

-0.90

 Hayward

-0.90

 Henderson

-0.88

POSITIVE LOGITS

 Polish

2.44

 Poland

2.37

ÅĤ

2.31

 Warsaw

2.26

 Poles

2.17

Pol

2.05

Å

1.67

Sz

1.62

 Lithuan

1.60

 Slovakia

1.57

Activations Density 0.270%

information related to Poland

references to geographic locations and their significance

No Comments

No Known Activations