INDEX

Explanations

names of countries and medical terms

oai_token-act-pair · gpt-3.5-turbo

keywords related to policies and organizations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

iona

-0.61

åĲ

-0.54

iku

-0.53

initial

-0.52

 Particip

-0.51

oppy

-0.51

 PsyNet

-0.50

Synopsis

-0.50

Events

-0.50

Deb

-0.50

POSITIVE LOGITS

foundland

0.78

etheless

0.78

 Conclusion

0.66

 Nept

0.65

anwhile

0.64

 ______

0.61

 queens

0.60

 wiser

0.59

CCC

0.58

tyr

0.58

Activations Density 1.379%

names of countries and medical terms

keywords related to policies and organizations

No Comments

No Known Activations