INDEX

Explanations

news headlines and articles related to politics and events

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

 telecommunications

-0.58

 disparate

-0.55

 surrog

-0.55

atform

-0.50

conservancy

-0.49

 Rodham

-0.48

 congressional

-0.48

athered

-0.48

zens

-0.48

 bona

-0.48

POSITIVE LOGITS

 ********************************

0.72

----------------------------------------------------------------

0.71

________________

0.69

 =================================================================

0.68

================================================================

0.67

 NOTE

0.67

________________________________________________________________

0.67

}}}

0.67

 ----------------------------------------------------------------

0.64

xx

0.64

Activations Density 3.681%

news headlines and articles related to politics and events

No Comments

No Known Activations