INDEX

Explanations

verbs in the imperative form, specifically focusing on instructions or suggestions

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_fs6144-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_6144

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

6,144

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

rones

-0.72

 Notting

-0.71

 Blitz

-0.67

Rab

-0.64

erity

-0.61

 cease

-0.61

 Desire

-0.60

phabet

-0.59

 Pear

-0.59

 Mans

-0.58

POSITIVE LOGITS

 traced

1.05

 viewed

1.05

 construed

1.00

 likened

1.00

 seen

1.00

 easily

0.99

 forgiven

0.98

 considered

0.91

 undone

0.90

 accommod

0.89

Activations Density 0.228%

verbs in the imperative form, specifically focusing on instructions or suggestions

No Comments

No Known Activations