INDEX

Explanations

mentions of specific names or terms

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

umer

-0.71

esis

-0.68

 reconc

-0.66

iments

-0.66

ervation

-0.65

 sacrific

-0.63

 occup

-0.63

eering

-0.63

emort

-0.63

terior

-0.63

POSITIVE LOGITS

forth

0.91

Berry

0.90

weed

0.88

lon

0.83

words

0.83

buck

0.82

mere

0.81

pee

0.79

down

0.79

adow

0.78

Activations Density 1.641%

mentions of specific names or terms

No Comments

No Known Activations