INDEX

Explanations

proper nouns related to various countries, people, organizations, and concepts

oai_token-act-pair · gpt-3.5-turbo

specific names and terms related to social issues, politics, and demographics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

DonaldTrump

-0.73

upper

-0.64

 Liga

-0.63

Ĥ¬

-0.63

mx

-0.61

ylum

-0.60

Ã¡s

-0.60

ibrary

-0.60

ELF

-0.59

arget

-0.59

POSITIVE LOGITS

 accordingly

1.15

 alike

1.13

 versa

1.06

 thereafter

0.97

 thereof

0.94

 therein

0.93

 thereto

0.93

 consequently

0.84

etheless

0.79

 consequ

0.78

Activations Density 0.675%

proper nouns related to various countries, people, organizations, and concepts

specific names and terms related to social issues, politics, and demographics

No Comments

No Known Activations