INDEX

Explanations

locations, especially shopping environments

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

param

-0.70

SHIP

-0.67

 resusc

-0.63

vernment

-0.63

Ake

-0.63

ioxide

-0.62

etting

-0.61

 exhaustion

-0.61

00200000

-0.60

 Instruments

-0.60

POSITIVE LOGITS

 malls

1.18

 Stores

1.18

 shoppers

1.17

 shelves

1.16

 stores

1.14

 storefront

1.13

 store

1.11

stores

1.08

 checkout

1.06

store

1.02

Activations Density 0.844%

locations, especially shopping environments

No Comments

No Known Activations