INDEX

Explanations

references to military and political events or entities, especially related to Russia

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_fs12288-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

12,288

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

 stub

-0.73

 itch

-0.67

rists

-0.64

 coughing

-0.64

 stray

-0.63

 symptoms

-0.63

 distinctions

-0.62

 smell

-0.62

 tantal

-0.61

 routine

-0.60

POSITIVE LOGITS

 Associates

0.70

 fame

0.70

ederation

0.69

 Strategic

0.68

 Operations

0.67

 Consulting

0.66

Lia

0.66

 Commando

0.65

iture

0.65

imedia

0.65

Activations Density 0.730%

references to military and political events or entities, especially related to Russia

No Comments

No Known Activations