INDEX

Explanations

language related to actions and events, often emphasizing progress or consequences

oai_token-act-pair · gpt-3.5-turbo

phrases indicating a cause-and-effect relationship

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 Nationals

-0.76

 Shutdown

-0.74

 Deadline

-0.74

 centr

-0.71

 Infrastructure

-0.70

Ĥİ

-0.69

FEC

-0.69

 Failure

-0.65

 USPS

-0.64

ETF

-0.64

POSITIVE LOGITS

 fascinated

1.16

 amazed

1.11

 admiration

1.10

 insights

1.03

 impressions

1.03

 amaz

1.03

 impressed

1.03

awe

1.02

 intrigued

0.99

 insight

0.99

Activations Density 0.628%

language related to actions and events, often emphasizing progress or consequences

phrases indicating a cause-and-effect relationship

No Comments

No Known Activations