INDEX

Explanations

words related to weight, weight loss, and physical attributes

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

Tul

-0.78

WP

-0.77

sky

-0.73

PBS

-0.72

 Philips

-0.70

 Solitaire

-0.69

 Noir

-0.69

 Bhar

-0.68

 Prairie

-0.68

 Ascend

-0.67

POSITIVE LOGITS

weight

1.25

 weights

1.14

 physique

1.10

 weight

1.07

Weight

1.06

weights

0.99

fat

0.96

 heaviest

0.93

 Weight

0.92

 stiffness

0.86

Activations Density 2.254%

words related to weight, weight loss, and physical attributes

No Comments

No Known Activations