INDEX

Explanations

phrases indicating a problematic situation or a small part hinting at a larger issue

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

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

BAT

-0.72

fy

-0.69

iors

-0.68

LY

-0.66

lords

-0.65

chev

-0.63

AMY

-0.62

cler

-0.61

OWN

-0.60

 lobb

-0.60

POSITIVE LOGITS

 iceberg

1.29

 scales

0.77

 scale

0.73

 proverbial

0.73

 rope

0.72

insula

0.70

 wedge

0.70

 mustard

0.68

 needles

0.62

 finger

0.61

Activations Density 0.073%

phrases indicating a problematic situation or a small part hinting at a larger issue

No Comments

No Known Activations

phrases indicating a problematic situation or a small part hinting at a larger issue

No Comments

No Known Activations