INDEX

Explanations

proper nouns and names of people, places, and organizations

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

»Ĵ

-0.70

å§«

-0.68

ousands

-0.68

acas

-0.65

iage

-0.65

ĥ

-0.59

wealth

-0.58

©¶æ

-0.58

seys

-0.58

sample

-0.57

POSITIVE LOGITS

 anyway

1.05

 rather

1.02

 anyways

0.91

 versus

0.90

 minus

0.87

!.

0.84

!!!

0.83

 insofar

0.81

!!!!

0.80

Activations Density 4.683%

proper nouns and names of people, places, and organizations

No Comments

No Known Activations