INDEX

Explanations

keyword "fix" and variations of the word

oai_token-act-pair · gpt-3.5-turbo

phrases related to correcting or addressing problems

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

ographic

-0.80

ographies

-0.75

 GOODMAN

-0.67

 sidx

-0.67

ography

-0.64

ographically

-0.63

ONSORED

-0.63

NetMessage

-0.62

arnaev

-0.61

istry

-0.61

POSITIVE LOGITS

ated

1.02

num

0.93

ate

0.87

ating

0.86

fix

0.81

tle

0.78

 bugs

0.78

up

0.77

ations

0.76

 broken

0.76

Activations Density 0.040%

keyword "fix" and variations of the word

phrases related to correcting or addressing problems

No Comments

No Known Activations