INDEX

Explanations

information related to research studies and statistics

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

inventoryQuantity

-1.01

 sooner

-0.82

redo

-0.82

 regrets

-0.77

 endif

-0.74

 anymore

-0.70

)}

-0.69

âĿ

-0.68

 rightfully

-0.67

 somew

-0.67

POSITIVE LOGITS

 simulated

1.25

 randomized

1.22

 randomly

1.18

 sampled

1.13

 samples

1.12

 participants

1.07

 sample

1.06

 volunteers

1.01

sample

1.00

 questionnaire

0.99

Activations Density 2.555%

information related to research studies and statistics

No Comments

No Known Activations