INDEX

Explanations

technical terms related to software and hardware

oai_token-act-pair · gpt-3.5-turbo

mentions of software and its specifications

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

natureconservancy

-0.51

 skelet

-0.47

 medic

-0.44

minist

-0.42

mel

-0.40

ap

-0.40

 willful

-0.39

 corpor

-0.38

 welf

-0.38

neg

-0.38

POSITIVE LOGITS

allery

0.63

ividual

0.58

glers

0.50

 "$:/

0.47

ãĤ¤ãĥĪ

0.46

ĨĴ

0.46

ĸļ

0.45

ãĤ¡

0.45

hower

0.45

UGE

0.45

Activations Density 0.507%

technical terms related to software and hardware

mentions of software and its specifications

No Comments

No Known Activations