INDEX

Explanations

information regarding numerical rankings or orders, particularly emphasizing items placed in second or second position

oai_token-act-pair · gpt-3.5-turbo

phrases indicating rankings or comparative positions

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

Ń·

-0.70

 formula

-0.66

 resume

-0.65

 Richards

-0.64

Ys

-0.62

 vapor

-0.62

 Shal

-0.60

 quote

-0.60

 Shutterstock

-0.59

POSITIVE LOGITS

generation

1.45

largest

1.42

highest

1.41

degree

1.35

tier

1.35

ranking

1.31

party

1.31

placed

1.30

ranked

1.29

place

1.27

Activations Density 0.028%

information regarding numerical rankings or orders, particularly emphasizing items placed in second or second position

phrases indicating rankings or comparative positions

No Comments

No Known Activations