INDEX

Explanations

phrases indicating uncertainty about location or position

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

iasis

-0.74

urated

-0.68

icer

-0.66

 Reef

-0.66

quel

-0.65

itri

-0.63

iger

-0.63

olas

-0.61

operation

-0.61

eton

-0.60

POSITIVE LOGITS

 else

1.43

abouts

1.26

 Else

1.12

Else

1.11

else

0.90

upon

0.89

 Near

0.88

 around

0.73

 along

0.73

around

0.73

Activations Density 0.069%

phrases indicating uncertainty about location or position

No Comments

No Known Activations