INDEX

Explanations

proper names

oai_token-act-pair · gpt-3.5-turbo

references to political figures and their affiliations

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

ĸļ

-1.00

 Jagu

-0.95

BMC

-0.82

 Slayer

-0.80

 experimental

-0.79

 metre

-0.78

 Milky

-0.78

 endeavour

-0.77

 Noct

-0.76

 Firefly

-0.76

POSITIVE LOGITS

Obama

1.57

OTUS

1.53

Romney

1.43

Trump

1.41

ghazi

1.39

Hillary

1.38

Clinton

1.37

Republicans

1.31

 Obama

1.27

Congress

1.27

Activations Density 0.380%

proper names

references to political figures and their affiliations

No Comments

No Known Activations