INDEX

Explanations

initialisms or acronyms related to years and organizations

oai_token-act-pair · gpt-3.5-turbo

fiscal years and related abbreviations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

houses

-0.73

 dotted

-0.70

 toget

-0.68

 Tasman

-0.64

Platform

-0.63

lain

-0.63

 barr

-0.63

 Cheong

-0.62

Companies

-0.61

 swast

-0.61

POSITIVE LOGITS

FY

1.13

FY

1.00

arie

0.83

ITS

0.75

ghan

0.74

iann

0.69

uly

0.69

CG

0.69

RS

0.69

adh

0.69

Activations Density 0.008%

initialisms or acronyms related to years and organizations

fiscal years and related abbreviations

No Comments

No Known Activations