INDEX

Explanations

phrases related to manufacturing and production processes

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

 recess

-0.70

 coward

-0.69

 Explan

-0.69

 adjourn

-0.69

secut

-0.67

Filename

-0.66

Leader

-0.66

anon

-0.66

dden

-0.66

 circumstance

-0.65

POSITIVE LOGITS

 products

1.31

 apparel

1.05

products

1.00

Products

0.98

 accessories

0.97

 merch

0.94

 services

0.94

 brands

0.92

services

0.92

 commodities

0.92

Activations Density 0.570%

phrases related to manufacturing and production processes

No Comments

No Known Activations