INDEX

Explanations

words related to research papers or studies, political processes, and sports analytics

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

 Combine

-0.62

aughs

-0.56

operation

-0.53

 Orchestra

-0.52

 Courier

-0.52

ipes

-0.52

ories

-0.51

}:

-0.51

 Cabin

-0.50

udeau

-0.50

POSITIVE LOGITS

 resembling

0.84

 whose

0.80

WithNo

0.79

 deemed

0.78

hots

0.78

outhern

0.72

 belonging

0.71

 pertaining

0.71

whose

0.70

 destined

0.69

Activations Density 3.196%

words related to research papers or studies, political processes, and sports analytics

No Comments

No Known Activations