INDEX

Explanations

references to legal or criminal activities

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_fs768-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

768

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 dissu

-0.76

lihood

-0.75

xual

-0.66

 overth

-0.65

 discourage

-0.64

 next

-0.63

 undecided

-0.63

 someday

-0.63

 outweigh

-0.63

 offline

-0.61

POSITIVE LOGITS

Posted

1.07

WASHINGTON

1.06

Contents

0.98

SHARE

0.97

ÂłÂłÂłÂłÂłÂłÂłÂł

0.95

SCP

0.93

LOS

0.92

Published

0.92

BBC

0.92

ccording

0.91

Activations Density 2.615%

references to legal or criminal activities

No Comments

No Known Activations