INDEX

Explanations

references to Angela Merkel

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

cffff

-0.84

nces

-0.78

ioned

-0.78

essee

-0.76

IFIED

-0.74

LOAD

-0.73

rolled

-0.73

ategory

-0.73

tery

-0.72

creen

-0.71

POSITIVE LOGITS

 Merkel

1.39

 Lans

0.90

lyn

0.86

 Marie

0.83

ï¸

0.80

 Thatcher

0.79

 Duck

0.77

 herself

0.76

llor

0.76

Anne

0.76

Activations Density 0.018%

references to Angela Merkel

No Comments

No Known Activations