INDEX

Explanations

numbers, symbols, and text encoding characters

oai_token-act-pair · gpt-3.5-turbo

currency symbols or related numeric representations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

cius

-0.74

BuyableInstoreAndOnline

-0.73

zona

-0.72

 Dragonbound

-0.66

angers

-0.66

 anat

-0.65

Downloadha

-0.63

ansas

-0.62

ItemImage

-0.61

ifice

-0.61

POSITIVE LOGITS

Ķ

1.10

¼

1.06

ł

1.04

¹

1.02

·

1.01

³

0.94

ħ

0.92

»

0.92

ª

0.91

âĶĢâĶĢâĶĢâĶĢâĶĢâĶĢâĶĢâĶĢ

0.90

Activations Density 0.007%

numbers, symbols, and text encoding characters

currency symbols or related numeric representations

No Comments

No Known Activations