INDEX

Explanations

positive or enthusiastic adjectives and phrases

oai_token-act-pair · gpt-3.5-turbo

expressions of strong positive sentiment or approval

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

 Garcia

-0.53

 fronts

-0.52

 Marvin

-0.50

 Greenwich

-0.49

unin

-0.49

 Nicarag

-0.48

 Kaufman

-0.48

 Pentagon

-0.48

 Shelby

-0.48

 Melvin

-0.47

POSITIVE LOGITS

âĢ

1.45

 âĢ

1.20

 âĢº

1.17

[/

1.14

 âľ

1.10

Â»

1.08

Â¨

1.08

 ðŁĳ

1.05

 ðŁ

1.04

âĹ

1.04

Activations Density 0.531%

positive or enthusiastic adjectives and phrases

expressions of strong positive sentiment or approval

No Comments

No Known Activations