INDEX

Explanations

phrases related to fleeing or running away

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

essee

-0.72

wcsstore

-0.66

inen

-0.65

iop

-0.64

eous

-0.64

Stretch

-0.63

alist

-0.63

sylv

-0.63

managed

-0.59

ãĤ¦ãĤ¹

-0.59

POSITIVE LOGITS

 altogether

0.91

 peacefully

0.91

 abruptly

0.88

 voluntarily

0.85

 overboard

0.78

 town

0.76

 unsc

0.76

 prematurely

0.74

 unnoticed

0.72

 quietly

0.67

Activations Density 0.813%

phrases related to fleeing or running away

No Comments

No Known Activations