INDEX

Explanations

headgear-related terms, specifically masks and helmets

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_fs12288-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

12,288

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

course

-0.87

inity

-0.74

ythm

-0.73

hower

-0.72

quart

-0.72

jun

-0.71

raltar

-0.71

tery

-0.67

ween

-0.66

pace

-0.66

POSITIVE LOGITS

 wearer

1.34

 worn

1.32

 goggles

0.96

 bearer

0.90

 adorned

0.87

 holder

0.87

 sleeve

0.86

 equipped

0.85

 mask

0.85

 sleeves

0.83

Activations Density 0.184%

headgear-related terms, specifically masks and helmets

No Comments

No Known Activations