INDEX

Explanations

mentions of political figures and events

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

 Collider

-0.73

scan

-0.73

Stream

-0.71

 Locations

-0.69

Origin

-0.68

 detector

-0.68

Depth

-0.67

sample

-0.67

pointers

-0.67

FIELD

-0.67

POSITIVE LOGITS

 dictatorship

1.28

 dictator

1.24

 regime

1.23

 decree

1.22

 appoint

1.22

 appointed

1.20

 administrations

1.19

 govern

1.17

 presidency

1.16

 administration

1.15

Activations Density 2.777%

mentions of political figures and events

No Comments

No Known Activations