INDEX

Explanations

references to economic and financial data, including statistics and figures

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

 bluff

-0.67

cha

-0.66

 fusion

-0.66

 masc

-0.63

 abras

-0.60

onite

-0.60

uclear

-0.60

 screws

-0.60

 barric

-0.60

creen

-0.59

POSITIVE LOGITS

 compared

0.94

 alone

0.93

-'

0.93

 dollars

0.86

 onwards

0.84

(%

0.83

adjusted

0.83

 Census

0.80

 Dollars

0.78

GDP

0.77

Activations Density 0.157%

references to economic and financial data, including statistics and figures

No Comments

No Known Activations