INDEX

Explanations

locations within a specific setting, particularly involving sitting or being situated in places

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

£ı

-0.78

cially

-0.71

²¾

-0.68

Ĥİ

-0.67

enthal

-0.64

 inst

-0.63

catentry

-0.63

ģ«

-0.61

 Logged

-0.61

 newsletter

-0.60

POSITIVE LOGITS

 chairs

0.90

 doorway

0.87

 seiz

0.87

 awaiting

0.81

 desks

0.81

cub

0.80

 porch

0.79

 waiting

0.79

 folding

0.78

 pedest

0.77

Activations Density 1.594%

locations within a specific setting, particularly involving sitting or being situated in places

No Comments

No Known Activations