INDEX

Explanations

documents, reports, memos, letters, and related administrative 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_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

animate

-0.68

 brightest

-0.66

 weddings

-0.65

cffff

-0.64

cale

-0.63

 distant

-0.61

udeau

-0.61

OPA

-0.59

 worship

-0.59

 unnatural

-0.59

POSITIVE LOGITS

 titled

1.37

 entitled

1.20

 detailing

1.10

 outlining

1.09

 published

0.95

 dated

0.92

velop

0.89

 summar

0.89

 documenting

0.88

 summarizes

0.85

Activations Density 2.273%

documents, reports, memos, letters, and related administrative terms

No Comments

No Known Activations