INDEX

Explanations

references to specific individuals or events in the 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_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

 accordingly

-0.81

 attent

-0.74

 awaited

-0.72

 recons

-0.71

 tracks

-0.71

 scaven

-0.71

 sooner

-0.70

 cues

-0.69

 prey

-0.69

 markings

-0.68

POSITIVE LOGITS

Flickr

1.35

Getty

1.33

AFP

1.25

Bloomberg

1.22

Reuters

1.15

Courtesy

1.14

Associated

1.09

Digital

1.02

CBC

1.01

Library

0.98

Activations Density 0.020%

references to specific individuals or events in the news

No Comments

No Known Activations