INDEX

Explanations

information related to news articles and current events

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

 defe

-0.76

ŃĶ

-0.73

 dips

-0.69

 compartment

-0.69

 landscape

-0.68

 utter

-0.66

 scenery

-0.66

 purse

-0.66

 pudding

-0.65

 tremend

-0.65

POSITIVE LOGITS

 Additionally

1.34

 However

1.29

 Meanwhile

1.28

 Specifically

1.27

 Previously

1.25

 According

1.25

 Instead

1.23

 Moreover

1.22

 Those

1.21

 Furthermore

1.19

Activations Density 2.051%

information related to news articles and current events

No Comments

No Known Activations