INDEX

Explanations

contractions

oai_token-act-pair · gpt-3.5-turbo

instances of negation or phrases indicating the absence of something

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 pyramid

-0.59

 warp

-0.56

 Mobil

-0.55

 Wilmington

-0.55

 chained

-0.54

 Lancaster

-0.54

 Doodle

-0.54

 cellphone

-0.54

 logger

-0.54

 Pyramid

-0.54

POSITIVE LOGITS

ï¸ı

1.04

Â£

0.94

Â¯Â¯

0.88

Â¢

0.86

âĪ

0.84

âĶĢâĶĢâĶĢâĶĢ

0.83

âĸł

0.82

Pg

0.82

Â§

0.81

âĻ

0.81

Activations Density 0.321%

contractions

instances of negation or phrases indicating the absence of something

No Comments

No Known Activations