INDEX

Explanations

discussions related to internet platforms or websites and information seeking behavior

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

 gains

-0.66

akable

-0.65

uries

-0.65

adan

-0.64

 Replay

-0.63

¿½

-0.62

 unle

-0.62

 depends

-0.62

ateral

-0.61

imble

-0.61

POSITIVE LOGITS

 myster

0.76

 hadn

0.74

 advertised

0.69

 Weird

0.69

Legend

0.67

 seemed

0.66

had

0.65

 advertisements

0.65

 Outbreak

0.62

 resembled

0.62

Activations Density 0.830%

discussions related to internet platforms or websites and information seeking behavior

No Comments

No Known Activations