INDEX

Explanations

the abbreviation "Vo" followed by a number

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

dress

-0.70

 foundations

-0.64

baugh

-0.63

 showers

-0.63

 iPads

-0.61

pillar

-0.61

taker

-0.60

 Scientist

-0.60

Reloaded

-0.58

 Journals

-0.58

POSITIVE LOGITS

VO

1.03

iced

0.97

zzo

0.96

irus

0.96

vernment

0.95

IP

0.94

Vo

0.94

icer

0.92

ile

0.92

ascular

0.90

Activations Density 0.014%

the abbreviation "Vo" followed by a number

No Comments

No Known Activations