INDEX

Explanations

references to programs, initiatives, and policies

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

 dilig

-0.68

pick

-0.66

 humidity

-0.65

ãĥ´

-0.64

 Cologne

-0.63

 statues

-0.63

uple

-0.63

ween

-0.62

 Robots

-0.62

Boss

-0.61

POSITIVE LOGITS

mable

2.01

matically

1.54

matic

1.53

mers

1.11

 Gutenberg

1.05

mes

0.96

ming

0.93

atically

0.91

atical

0.89

zee

0.86

Activations Density 0.572%

references to programs, initiatives, and policies

No Comments

No Known Activations