INDEX

Explanations

terms related to technical specifications and features

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

nikov

-0.71

Deal

-0.70

>[

-0.68

pees

-0.67

enfranch

-0.64

pell

-0.64

gest

-0.63

hews

-0.63

zb

-0.63

bucks

-0.62

POSITIVE LOGITS

rogens

1.11

romeda

1.09

 thence

1.09

 consequently

1.08

rogen

1.07

 hence

1.06

 thus

1.02

 therefore

1.00

 subsequently

0.97

 consequ

0.95

Activations Density 1.550%

terms related to technical specifications and features

No Comments

No Known Activations