INDEX

Explanations

words related to experiments or research activities

oai_token-act-pair · gpt-3.5-turbo

occurrences of the substring "erm."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 Seat

-0.72

 bowls

-0.63

 feast

-0.62

 sights

-0.60

Ike

-0.59

 temples

-0.59

 principals

-0.59

 saturated

-0.59

åĪ

-0.58

 kitchens

-0.58

POSITIVE LOGITS

erm

4.75

ern

1.51

erb

1.51

erman

1.45

erg

1.36

err

1.35

erk

1.29

eri

1.11

erry

1.10

ert

1.02

Activations Density 0.011%

words related to experiments or research activities

occurrences of the substring "erm."

No Comments

No Known Activations