INDEX

Explanations

numeric values associated with quantities and measurements

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_fs768-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

768

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

warts

-0.69

venge

-0.65

ufact

-0.63

bara

-0.62

iqu

-0.61

agall

-0.61

comes

-0.61

nam

-0.60

sole

-0.60

Discuss

-0.59

POSITIVE LOGITS

 pounds

1.22

lbs

1.17

 hectares

1.14

 acres

1.12

 gallons

1.10

 dollars

1.09

lbs

1.07

lb

1.07

 miles

1.05

 tonnes

1.04

Activations Density 2.190%

numeric values associated with quantities and measurements

No Comments

No Known Activations