INDEX

Explanations

names of companies and pharmaceutical 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_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

olated

-0.74

veyard

-0.73

qqa

-0.72

 "$:/

-0.69

liest

-0.66

idth

-0.66

 filibuster

-0.65

indu

-0.64

idency

-0.64

geries

-0.64

POSITIVE LOGITS

 Corporation

1.35

 Corp

1.29

 Holdings

1.27

 subsidiary

1.21

CEO

1.16

 Industries

1.16

 Technologies

1.16

 executives

1.15

Ltd

1.14

 subsidiaries

1.12

Activations Density 1.819%

names of companies and pharmaceutical terms

No Comments

No Known Activations