INDEX

Explanations

technical information related to scientific experiments and discoveries

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

 grievance

-0.89

iquette

-0.84

 entitlement

-0.81

 treasury

-0.79

 patriotism

-0.79

 apost

-0.79

 estate

-0.78

 discretionary

-0.78

street

-0.76

 tier

-0.75

POSITIVE LOGITS

 Researchers

1.67

 Scientists

1.63

Scientists

1.47

Researchers

1.45

 Canaver

1.33

 Moreover

1.14

 Furthermore

1.12

 Experts

1.10

 Investigators

1.09

 Molecular

1.08

Activations Density 0.348%

technical information related to scientific experiments and discoveries

No Comments

No Known Activations