INDEX

Explanations

Twitter links with high activity levels (towards the end of the link)

oai_token-act-pair · gpt-3.5-turbo

occurrences of the letter 't'

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

ĪĴ

-0.88

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

-0.74

©¶æ

-0.69

 exha

-0.68

ĻĤ

-0.64

 Hitman

-0.62

²¾

-0.61

 neighb

-0.60

Ĥİ

-0.60

 Arri

-0.59

POSITIVE LOGITS

ribune

1.03

rell

0.92

ruly

0.89

ibia

0.89

oxic

0.88

itans

0.88

ricks

0.88

ravis

0.87

akedown

0.86

ebin

0.86

Activations Density 0.013%

Twitter links with high activity levels (towards the end of the link)

occurrences of the letter 't'

No Comments

No Known Activations