INDEX

Explanations

mentions of military actions, particularly airstrikes and attacks

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

 Wealth

-0.79

MpServer

-0.76

coin

-0.76

FINE

-0.74

 Coin

-0.71

resents

-0.70

Fill

-0.67

 Edison

-0.66

ħĭ

-0.65

Merit

-0.65

POSITIVE LOGITS

 airstrikes

1.07

 targeting

1.07

 airstrike

1.04

 bombardment

1.01

 bombing

0.99

 aimed

0.98

 helicopters

0.96

 targeted

0.96

 raids

0.95

 bombed

0.91

Activations Density 0.076%

mentions of military actions, particularly airstrikes and attacks

No Comments

No Known Activations