INDEX

Explanations

dates or 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

 Mariners

-0.66

 NAACP

-0.62

 priesthood

-0.60

 Priest

-0.58

 Engineers

-0.57

 Ment

-0.57

 Zeit

-0.57

 Mole

-0.56

HC

-0.56

 Episcopal

-0.55

POSITIVE LOGITS

 beforehand

0.87

 dearly

0.82

 throughout

0.80

 during

0.77

.:

0.77

abouts

0.73

 before

0.72

.?

0.72

 whilst

0.70

.</

0.70

Activations Density 0.914%

dates or events

No Comments

No Known Activations