INDEX

Explanations

descriptions related to various events and situations

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

)"

-0.80

etc

-0.71

arta

-0.70

 However

-0.70

alde

-0.68

 Whilst

-0.67

[+

-0.66

 fixme

-0.66

upon

-0.65

 UNCLASSIFIED

-0.65

POSITIVE LOGITS

 nonetheless

1.16

etheless

1.04

 downright

0.91

 awfully

0.79

 decidedly

0.74

 elusive

0.74

 insidious

0.73

 remarkably

0.72

 lurking

0.71

 nevertheless

0.71

Activations Density 4.508%

descriptions related to various events and situations

No Comments

No Known Activations