INDEX

Explanations

terms related to non-fiction literature

oai_token-act-pair · gpt-3.5-turbo

references to nonfiction and fiction genres

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 reciproc

-0.66

arov

-0.65

Ķ

-0.65

VP

-0.64

gars

-0.62

CHR

-0.62

bred

-0.61

Accessory

-0.60

 Michaels

-0.60

 McMahon

-0.57

POSITIVE LOGITS

fiction

1.31

 Fiction

0.96

 fiction

0.90

acters

0.78

ipeg

0.75

DragonMagazine

0.74

ussion

0.74

ocity

0.73

aceous

0.72

 guiActiveUn

0.72

Activations Density 0.008%

terms related to non-fiction literature

references to nonfiction and fiction genres

No Comments

No Known Activations