INDEX

Explanations

information related to companies, products, and statistics

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

conservancy

-1.00

otin

-0.76

bably

-0.74

ecause

-0.72

 worrying

-0.72

 speeding

-0.71

zens

-0.70

 destro

-0.70

 cultivating

-0.70

 bolst

-0.69

POSITIVE LOGITS

 Description

1.14

 Location

1.14

 Released

1.09

 Includes

1.08

 Available

1.06

TBA

1.05

 Contains

1.04

 Date

1.03

 Retrieved

1.03

 âĵĺ

1.02

Activations Density 5.938%

information related to companies, products, and statistics

No Comments

No Known Activations