INDEX

Explanations

negative actions and events related to terrorism and security issues

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

 Remastered

-0.73

AVG

-0.70

 plates

-0.67

 coat

-0.67

 refresh

-0.66

 props

-0.65

 bumps

-0.64

OPLE

-0.64

 pancakes

-0.63

 scratches

-0.63

POSITIVE LOGITS

Qaeda

1.41

Qaida

1.35

Assad

1.16

raq

1.10

Monitor

1.07

Bay

1.04

Islam

1.02

Az

1.01

Nusra

0.97

Lat

0.96

Activations Density 0.026%

negative actions and events related to terrorism and security issues

No Comments

No Known Activations