INDEX

Explanations

phrases related to political systems, such as the Electoral College and voter turnout

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

shirt

-0.71

 helicop

-0.70

TG

-0.67

Train

-0.65

SHIP

-0.64

owship

-0.64

 Vessel

-0.64

Tweet

-0.63

Newsletter

-0.63

lass

-0.62

POSITIVE LOGITS

 origins

1.21

 significance

1.17

 resemblance

1.14

 implications

1.13

 similarities

1.12

 complexity

1.07

 relevance

1.07

 drawbacks

1.06

 similarity

1.03

 resonance

1.02

Activations Density 1.698%

phrases related to political systems, such as the Electoral College and voter turnout

No Comments

No Known Activations