INDEX

Explanations

specific terms and phrases related to legal matters and technological processes

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Glas

-0.66

 aback

-0.63

fitting

-0.62

anmar

-0.61

 thanking

-0.60

olk

-0.60

 Lank

-0.60

 Franch

-0.57

 Canaver

-0.57

 Gram

-0.56

POSITIVE LOGITS

 exists

1.28

 disappears

1.18

 lasts

1.15

 expires

1.11

 fails

1.09

 survives

1.07

 exceeds

1.06

 contains

1.06

 becomes

1.05

 shines

1.04

Activations Density 0.471%

specific terms and phrases related to legal matters and technological processes

No Comments

No Known Activations