INDEX

Explanations

ordinal numbers in a specific context

oai_token-act-pair · gpt-3.5-turbo

phrases indicating a ranking or classification, particularly the term "second."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 Arabian

-0.75

 Richards

-0.69

 Paradox

-0.69

pload

-0.68

ãĤ¡

-0.66

 Bullets

-0.65

AVG

-0.65

 DRAG

-0.64

 proxies

-0.62

 Timbers

-0.62

POSITIVE LOGITS

generation

1.21

degree

1.19

eyed

1.14

level

1.11

division

1.10

dimensional

1.08

year

1.08

century

1.07

chance

1.06

party

1.06

Activations Density 0.109%

ordinal numbers in a specific context

phrases indicating a ranking or classification, particularly the term "second."

No Comments

No Known Activations