INDEX

Explanations

technical references related to specific software features or bugs

oai_token-act-pair · gpt-3.5-turbo

instances of the term "Ex" followed by various numerical values or indicators

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

taboola

-0.80

Kuh

-0.75

oÄŁ

-0.67

yard

-0.65

MENTS

-0.64

Ń·

-0.61

riter

-0.61

Hoy

-0.60

 crawl

-0.60

 loudspe

-0.59

POSITIVE LOGITS

haust

1.15

terior

1.12

posure

1.12

odus

1.08

Ex

1.03

ternally

1.02

ercise

1.00

cerpt

0.96

asper

0.94

change

0.93

Activations Density 0.008%

technical references related to specific software features or bugs

instances of the term "Ex" followed by various numerical values or indicators

No Comments

No Known Activations