INDEX

Explanations

specific entities such as locations, organizations, and individuals

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

}.

-0.67

âĢ¢âĢ¢

-0.65

ä»£

-0.64

ĸļ

-0.63

*.

-0.59

>.

-0.58

_.

-0.58

*.

-0.56

 basis

-0.56

ãĢĤ

-0.56

POSITIVE LOGITS

coin

0.65

Coin

0.64

justice

0.61

anta

0.59

ratom

0.58

 officials

0.58

Stream

0.58

Sec

0.56

Tel

0.56

 Brewing

0.55

Activations Density 0.579%

specific entities such as locations, organizations, and individuals

No Comments

No Known Activations