INDEX

Explanations

adjectives related to exaggeration or comparison

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

ashtra

-0.80

afety

-0.78

iard

-0.69

ascript

-0.68

utical

-0.66

tery

-0.66

anmar

-0.65

Ĥİ

-0.65

gans

-0.65

essor

-0.64

POSITIVE LOGITS

 Savior

0.75

Irish

0.74

 Gleaming

0.71

Sax

0.71

ellation

0.70

Eye

0.69

icket

0.69

 Strip

0.68

Letter

0.68

Led

0.67

Activations Density 0.065%

adjectives related to exaggeration or comparison

No Comments

No Known Activations