INDEX

Explanations

mentions of different colors and faiths, particularly in a social context

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

Quantity

-0.71

alloc

-0.69

EVA

-0.67

 sensations

-0.66

washer

-0.65

Examples

-0.64

Deal

-0.63

Prosecut

-0.62

ebin

-0.61

Features

-0.61

POSITIVE LOGITS

 whom

1.27

 attendance

0.76

pires

0.75

who

0.71

whose

0.70

hran

0.69

 backgrounds

0.69

pired

0.69

 alike

0.68

 professions

0.68

Activations Density 1.436%

mentions of different colors and faiths, particularly in a social context

No Comments

No Known Activations