INDEX

Explanations

words related to law enforcement and specific police-related terms

oai_token-act-pair · gpt-3.5-turbo

mentions of specific locations and fried food items

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 induce

-0.78

 humid

-0.69

TN

-0.66

isp

-0.64

 cessation

-0.64

 imper

-0.62

 tele

-0.62

 tribal

-0.61

 conservative

-0.61

 freeze

-0.60

POSITIVE LOGITS

 Yard

4.60

 Yards

1.99

yard

1.27

 Alley

1.15

 Rack

1.13

 Garage

1.13

 Shard

1.07

Row

1.02

yards

1.02

 Loft

0.97

Activations Density 0.018%

words related to law enforcement and specific police-related terms

mentions of specific locations and fried food items

No Comments

No Known Activations