INDEX

Explanations

the word "California"

oai_token-act-pair · gpt-3.5-turbo

references to California

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

Netflix

-0.80

AIDS

-0.75

Match

-0.71

 Hoff

-0.71

listed

-0.70

 Pirate

-0.70

jobs

-0.69

 Turing

-0.67

 Hulu

-0.66

anded

-0.65

POSITIVE LOGITS

cal

4.05

Cal

1.55

cal

1.54

CAL

1.51

Cal

1.28

 calc

1.04

sul

1.02

tem

0.98

cu

0.97

cam

0.95

Activations Density 0.015%

the word "California"

references to California

No Comments

No Known Activations