INDEX

Explanations

scientific publications and research-related terms

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

lou

-0.70

chwitz

-0.69

rower

-0.68

hog

-0.68

haul

-0.68

 towers

-0.66

bnb

-0.65

assian

-0.65

vous

-0.65

landish

-0.64

POSITIVE LOGITS

 Pediatrics

1.05

 Psychiatry

1.05

 Perspect

1.03

 Supplement

1.00

 Biol

0.99

 Biology

0.96

 Neuroscience

0.95

 Journal

0.94

 PLoS

0.93

doi

0.93

Activations Density 0.068%

scientific publications and research-related terms

No Comments

No Known Activations