INDEX

Explanations

words related to instructions, maintenance, comparison, and clarification

oai_token-act-pair · gpt-3.5-turbo

terms related to formal processes and evaluations

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

olated

-0.68

apa

-0.68

anon

-0.65

amaru

-0.63

former

-0.63

habi

-0.62

mare

-0.59

ivan

-0.58

inese

-0.56

gall

-0.56

POSITIVE LOGITS

 purposes

2.33

 sake

1.88

 reasons

1.57

 purpose

1.23

ummies

1.06

 reason

0.97

 Reasons

0.91

 occasions

0.82

 emergencies

0.80

 eternity

0.78

Activations Density 0.373%

words related to instructions, maintenance, comparison, and clarification

terms related to formal processes and evaluations

No Comments

No Known Activations