INDEX

Explanations

technical terms and numerical values, potentially related to scientific or factual information

oai_token-act-pair · gpt-3.5-turbo

words related to factual information and measurement

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

Cash

-0.66

 Cadillac

-0.59

ieri

-0.58

audi

-0.57

 Cash

-0.57

-0.56

\'

-0.55

Tec

-0.55

ANA

-0.53

ascal

-0.53

POSITIVE LOGITS

pedia

0.61

 experien

0.56

 reperto

0.53

 pals

0.52

watching

0.52

licted

0.50

 nomine

0.50

gered

0.50

worthiness

0.50

 recre

0.50

Activations Density 1.204%

technical terms and numerical values, potentially related to scientific or factual information

words related to factual information and measurement

No Comments

No Known Activations