INDEX

Explanations

mentions of publications, media outlets, and specific types of interviews

oai_token-act-pair · gpt-3.5-turbo

mentions of interviews and media organizations

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

cause

-0.69

$.

-0.63

 whichever

-0.61

laughs

-0.60

%).

-0.58

"},"

-0.58

"!

-0.57

.''.

-0.56

._

-0.56

%;

-0.56

POSITIVE LOGITS

 yesterday

1.14

 Tuesday

1.00

 Thursday

1.00

 Wednesday

0.99

 earlier

0.95

 Monday

0.94

 Friday

0.92

terday

0.89

 last

0.89

 shortly

0.83

Activations Density 0.205%

mentions of publications, media outlets, and specific types of interviews

mentions of interviews and media organizations

No Comments

No Known Activations