INDEX

Explanations

season numbers within text

oai_token-act-pair · gpt-3.5-turbo

references to television show seasons

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

ntil

-0.81

artisan

-0.77

Ney

-0.73

ulic

-0.68

imir

-0.67

iasis

-0.66

uggage

-0.65

roleum

-0.64

jab

-0.64

orget

-0.64

POSITIVE LOGITS

 finale

1.46

 premiere

1.36

 premie

1.24

 premiered

1.17

 episode

1.16

 episodes

1.13

 installments

1.12

 airing

1.11

episode

1.10

 Episode

1.07

Activations Density 0.133%

season numbers within text

references to television show seasons

No Comments

No Known Activations