INDEX

Explanations

proper nouns or names of people, places, or organizations

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

 anytime

-0.65

ario

-0.65

ysc

-0.62

 translates

-0.61

irez

-0.61

rification

-0.60

culosis

-0.59

ober

-0.58

 introduction

-0.58

 miscarriage

-0.58

POSITIVE LOGITS

 those

0.84

Ĳ

0.72

those

0.72

 whom

0.69

ĪĴ

0.68

addons

0.66

 innumerable

0.66

 them

0.66

 casualties

0.66

Sources

0.65

Activations Density 0.027%

proper nouns or names of people, places, or organizations

No Comments

No Known Activations