INDEX

Explanations

amount or quantity-related statements

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_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

agree

-0.75

always

-0.70

Ec

-0.66

 endeavors

-0.65

anism

-0.65

adding

-0.65

Events

-0.65

 aloud

-0.64

Iss

-0.64

Therefore

-0.62

POSITIVE LOGITS

 foothold

1.24

 glimpse

1.18

hold

1.13

 chance

1.01

lot

0.99

 haircut

0.88

 bead

0.85

 taste

0.85

 clue

0.83

 traction

0.82

Activations Density 0.102%

amount or quantity-related statements

No Comments

No Known Activations