INDEX

Explanations

information related to news organizations, specific journalists, and articles they have written

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

ranged

-0.75

Qi

-0.68

inho

-0.67

Aki

-0.66

ayne

-0.65

cific

-0.64

 Berm

-0.64

 Garr

-0.62

Ryu

-0.61

 Luigi

-0.61

POSITIVE LOGITS

 publishes

1.00

 broadcasts

0.97

 publishing

0.85

 Newsp

0.84

 broadcasting

0.83

 feeds

0.82

 publisher

0.82

 journalism

0.82

 subscriptions

0.81

 viewership

0.81

Activations Density 0.275%

information related to news organizations, specific journalists, and articles they have written

No Comments

No Known Activations