INDEX

Explanations

technical specifications and performance details

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

Deal

-0.75

dq

-0.72

enfranch

-0.68

umo

-0.67

phot

-0.63

uma

-0.62

inces

-0.61

.*

-0.61

 distingu

-0.60

jah

-0.60

POSITIVE LOGITS

rogens

1.26

romeda

1.18

rogen

1.16

ERSON

1.00

 thence

0.90

rew

0.90

 consequently

0.88

 Beyond

0.84

 therefore

0.84

 subsequently

0.81

Activations Density 1.044%

technical specifications and performance details

No Comments

No Known Activations