INDEX

Explanations

phrases indicating determination or decision

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.10.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

aples

-0.67

ernels

-0.66

WIND

-0.61

archives

-0.60

ilers

-0.59

Kis

-0.58

wine

-0.58

heed

-0.57

asus

-0.57

illi

-0.57

POSITIVE LOGITS

 anywhere

1.32

 anymore

1.10

nor

0.90

 slightest

0.87

 anytime

0.86

any

0.84

 overboard

0.84

 whatsoever

0.81

 unnoticed

0.80

 lightly

0.77

Activations Density 0.048%

phrases indicating determination or decision

No Comments

No Known Activations