INDEX

Explanations

phrases related to introspection and analysis of different 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_fs12288-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

12,288

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

orb

-0.62

EStreamFrame

-0.61

theless

-0.60

Cele

-0.60

CENT

-0.59

cffff

-0.59

ä¹

-0.58

âĹ¼

-0.57

accompan

-0.55

cape

-0.55

POSITIVE LOGITS

ahead

0.97

 favorably

0.76

 suspic

0.73

 forward

0.72

izons

0.69

 alike

0.68

 into

0.68

ãĤ¶

0.67

behind

0.67

 upon

0.67

Activations Density 0.087%

phrases related to introspection and analysis of different situations

No Comments

No Known Activations