INDEX

Explanations

mentions of officials conducting investigations or confirming information in news articles

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

;;;;;;;;;;;;

-0.71

Interstitial

-0.70

 Seat

-0.69

ãĥĸ

-0.66

series

-0.64

Translation

-0.64

gets

-0.63

 Oval

-0.62

 Series

-0.62

bearing

-0.62

POSITIVE LOGITS

 speculate

1.29

 allege

1.08

 contend

1.03

 speculated

1.02

 estimate

1.02

 believe

1.01

say

1.00

 wondered

0.99

 anticipate

0.96

 theor

0.95

Activations Density 2.882%

mentions of officials conducting investigations or confirming information in news articles

No Comments

No Known Activations