INDEX

Explanations

informational content related to specific products or services

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_fs12288-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

12,288

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

usercontent

-0.84

chest

-0.77

OGR

-0.68

opic

-0.67

emo

-0.66

ndra

-0.64

inent

-0.64

atican

-0.64

ELY

-0.63

oters

-0.62

POSITIVE LOGITS

how

0.91

 pricing

0.84

 specifics

0.80

 topics

0.79

 each

0.79

 upcoming

0.79

 whats

0.79

why

0.77

 what

0.74

 licensing

0.73

Activations Density 0.273%

informational content related to specific products or services

No Comments

No Known Activations