INDEX

Explanations

references to economic terms and analysis

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

vale

-0.76

endo

-0.74

mie

-0.72

REDACTED

-0.71

ificent

-0.70

ãĤĵ

-0.67

ISION

-0.67

ocese

-0.66

ovich

-0.65

iquette

-0.63

POSITIVE LOGITS

 namely

1.06

 whereby

1.02

 including

1.01

aka

1.00

 especially

0.98

 wherein

0.97

 particularly

0.97

particularly

0.97

Downloadha

0.95

especially

0.95

Activations Density 0.401%

references to economic terms and analysis

No Comments

No Known Activations