INDEX

Explanations

comparative descriptions using the word "more"

oai_token-act-pair · gpt-3.5-turbo

the phrase "for more," indicating additional information or content

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

Enlarge

-0.82

pload

-0.79

ader

-0.77

adr

-0.73

Cro

-0.70

xtap

-0.70

oder

-0.69

Fram

-0.68

ption

-0.68

Disk

-0.67

POSITIVE LOGITS

 info

1.05

 information

1.04

 details

0.95

 than

0.88

 detailed

0.86

 interesting

0.86

 informative

0.85

 insight

0.85

 INFORMATION

0.80

 insights

0.80

Activations Density 0.032%

comparative descriptions using the word "more"

the phrase "for more," indicating additional information or content

No Comments

No Known Activations