INDEX

Explanations

phrases that indicate attribution or reference to sources

oai_token-act-pair · gpt-4o-mini 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

obyl

-0.80

eg

-0.67

otin

-0.67

igmat

-0.66

ãĤ¨ãĥ«

-0.65

ÃĥÃĤ

-0.65

affe

-0.62

IFT

-0.61

 squared

-0.61

cair

-0.60

POSITIVE LOGITS

 sources

1.03

Sources

0.75

SOURCE

0.72

 Reuters

0.69

WS

0.66

 Sources

0.66

SPONSORED

0.65

axter

0.63

CNS

0.63

 citing

0.62

Activations Density 0.038%

phrases that indicate attribution or reference to sources

No Comments

No Known Activations