INDEX

Explanations

references to legal or regulatory issues, as well as political and economic news

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_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

 predecessor

-0.73

xxxx

-0.67

ONSORED

-0.65

isSpecialOrderable

-0.64

washer

-0.63

izable

-0.62

 belongs

-0.62

66666666

-0.61

TX

-0.61

 signifies

-0.61

POSITIVE LOGITS

 clam

1.14

 alike

1.13

 routinely

1.06

 flock

1.00

 pressured

0.97

 understandably

0.96

 collectively

0.96

 increasingly

0.96

 rejoice

0.91

ateurs

0.91

Activations Density 2.323%

references to legal or regulatory issues, as well as political and economic news

No Comments

No Known Activations