INDEX

Explanations

mentions of values related to respect and integrity

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

 Lans

-0.93

 concoct

-0.75

 helicop

-0.68

 paran

-0.67

 dreamed

-0.67

NetMessage

-0.66

otine

-0.65

 nerv

-0.65

uzzle

-0.64

 enthusi

-0.64

POSITIVE LOGITS

ability

1.47

ably

1.35

fully

1.04

able

1.03

abilities

1.01

ility

0.93

ful

0.91

ibly

0.90

amental

0.88

FUL

0.88

Activations Density 0.028%

mentions of values related to respect and integrity

No Comments

No Known Activations