INDEX

Explanations

references to the political organization Hamas

oai_token-act-pair · gpt-3.5-turbo

mentions of "Hamas."

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

 target

-0.66

 level

-0.63

 Prof

-0.63

 tick

-0.62

 Reading

-0.62

 Well

-0.58

shy

-0.58

Dri

-0.57

 charging

-0.57

 Curious

-0.57

POSITIVE LOGITS

amas

4.68

ama

2.11

amic

1.44

AMA

1.41

anan

1.32

aman

1.25

amac

1.21

amia

1.20

oms

1.13

ams

1.11

Activations Density 0.012%

references to the political organization Hamas

mentions of "Hamas."

No Comments

No Known Activations