INDEX

Explanations

information related to research, scientific studies, medical errors, justice system statistics, elections, and public health reports

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

 fame

-0.66

 reckoning

-0.66

deen

-0.60

liness

-0.60

Hack

-0.58

ãĥ¼ãĥ«

-0.58

Reviewed

-0.58

timer

-0.58

claimer

-0.57

inventoryQuantity

-0.57

POSITIVE LOGITS

 percent

1.10

 million

1.03

 consecutive

0.98

 additional

0.90

 dozen

0.88

 fewer

0.87

%"

0.86

0.85

 instances

0.83

Activations Density 1.635%

information related to research, scientific studies, medical errors, justice system statistics, elections, and public health reports

No Comments

No Known Activations