INDEX

Explanations

words related to specific technology products

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

ozyg

-0.81

glers

-0.76

hood

-0.75

bush

-0.74

jar

-0.71

doms

-0.70

 Fenrir

-0.69

chens

-0.63

lihood

-0.62

tons

-0.62

POSITIVE LOGITS

CEPT

1.08

ORN

1.04

oustic

0.99

HE

0.95

APTER

0.92

oust

0.92

olor

0.91

SI

0.91

ATA

0.90

ME

0.88

Activations Density 0.012%

words related to specific technology products

No Comments

No Known Activations