INDEX

Explanations

phrases related to legal or official documentation

oai_token-act-pair · gpt-3.5-turbo

legal terminology and references to official processes

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

âĢ

-0.99

His

-0.79

 ÂŃ

-0.78

elight

-0.75

)</

-0.73

âĹ

-0.72

âĸº

-0.72

"},{"

-0.72

.","

-0.71

POSITIVE LOGITS

TWO

0.77

 EVERY

0.76

ALL

0.73

ANY

0.72

 folks

0.70

 ONLY

0.68

 negro

0.67

DOD

0.67

 THIS

0.66

 THAT

0.65

Activations Density 1.081%

phrases related to legal or official documentation

legal terminology and references to official processes

No Comments

No Known Activations