INDEX

Explanations

words related to discussing or addressing specific topics or issues

oai_token-act-pair · gpt-3.5-turbo

phrases indicative of oversight, neglect, or lack of attention in discussions

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

yssey

-0.75

à¼

-0.72

Torrent

-0.71

STAR

-0.71

KO

-0.70

unte

-0.69

NT

-0.66

vic

-0.66

spawn

-0.66

Spawn

-0.64

POSITIVE LOGITS

 topics

1.16

 anecdotes

1.01

 specifics

1.00

 issues

0.96

 facts

0.89

 methodological

0.88

 whether

0.88

 aspects

0.88

 controversies

0.85

 subjects

0.85

Activations Density 0.408%

words related to discussing or addressing specific topics or issues

phrases indicative of oversight, neglect, or lack of attention in discussions

No Comments

No Known Activations