INDEX

Explanations

information related to political events, especially pertaining to leadership changes and legal actions

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

 Collider

-0.83

Sym

-0.76

 Cosponsors

-0.73

imentary

-0.71

rawdownloadcloneembedreportprint

-0.69

 detector

-0.69

 bandwidth

-0.68

Drop

-0.68

Spot

-0.67

Recommend

-0.67

POSITIVE LOGITS

 dictatorship

1.35

 ousted

1.22

 democratically

1.18

 dictator

1.18

 overth

1.17

 regime

1.17

 monarchy

1.16

 rulers

1.14

 decree

1.12

 govern

1.10

Activations Density 4.221%

information related to political events, especially pertaining to leadership changes and legal actions

No Comments

No Known Activations