INDEX

Explanations

actions related to protecting, forgiving, warning, and returning

oai_token-act-pair · gpt-3.5-turbo

themes of forgiveness and redemption in actions

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

utterstock

-0.86

Twitter

-0.84

Republican

-0.81

earchers

-0.80

HOU

-0.76

Federal

-0.76

Reviewer

-0.76

FOX

-0.74

urches

-0.74

Michigan

-0.73

POSITIVE LOGITS

 Luffy

1.15

his

1.13

him

1.05

Yug

1.04

 Amon

0.99

 Zoro

0.99

Iz

0.98

 Izan

0.97

 hers

0.97

 Jinn

0.96

Activations Density 0.321%

actions related to protecting, forgiving, warning, and returning

themes of forgiveness and redemption in actions

No Comments

No Known Activations