INDEX

Explanations

contractions containing a negative assertion

oai_token-act-pair · gpt-3.5-turbo

the repetition of the word "doesn't."

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

 hitherto

-0.56

 present

-0.55

wit

-0.52

 departing

-0.52

 respective

-0.52

 grave

-0.51

 occasion

-0.51

 considerable

-0.51

 approximately

-0.50

ivery

-0.50

POSITIVE LOGITS

 doesn

2.90

doesn

2.49

isn

2.39

 doesnt

2.32

 hasn

2.30

 Doesn

2.18

 didn

2.02

 wasn

2.00

 shouldn

1.94

 DOES

1.94

Activations Density 0.031%

contractions containing a negative assertion

the repetition of the word "doesn't."

No Comments

No Known Activations