INDEX

Explanations

references to military-related concepts like veterans, defense, and military intervention

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

âĶģ

-0.88

 Seller

-0.78

flush

-0.73

Lot

-0.72

apest

-0.72

Publisher

-0.69

à

-0.67

 Rowling

-0.67

luck

-0.67

nob

-0.64

POSITIVE LOGITS

 fatig

1.31

 stationed

1.26

 barracks

1.24

 colonel

1.23

 commanders

1.12

 Corps

1.10

 commander

1.09

 garrison

1.05

 corps

1.01

 personnel

0.98

Activations Density 2.114%

references to military-related concepts like veterans, defense, and military intervention

No Comments

No Known Activations