INDEX

Explanations

mentions of political figures, specifically presidents

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_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

lighting

-0.73

essen

-0.70

lot

-0.69

ensions

-0.65

MpServer

-0.64

vana

-0.61

doors

-0.60

ratom

-0.60

âĵĺ

-0.60

 hitch

-0.59

POSITIVE LOGITS

 Barack

1.13

ially

1.09

 Obama

1.02

clinton

1.02

ial

0.93

 Lyndon

0.92

 Recep

0.87

 Donald

0.87

IAL

0.83

 Bashar

0.83

Activations Density 1.174%

mentions of political figures, specifically presidents

No Comments

No Known Activations