INDEX

Explanations

numerical estimates or amounts

oai_token-act-pair · gpt-3.5-turbo

phrases indicating estimations or approximations of numbers

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

ĪĴ

-0.86

neys

-0.80

woods

-0.77

GU

-0.77

ription

-0.75

uits

-0.74

gob

-0.71

ighting

-0.69

feeding

-0.69

okÃ©

-0.68

POSITIVE LOGITS

 amount

0.84

 number

0.80

 seventy

0.80

 âĤ¬

0.78

 sixty

0.78

0.75

 eighty

0.74

 lifespan

0.74

Activations Density 0.043%

numerical estimates or amounts

phrases indicating estimations or approximations of numbers

No Comments

No Known Activations