INDEX

Explanations

terms related to investigation and evidence, including statements regarding the absence or presence of evidence, concrete knowledge, and findings from investigations

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

picture

-0.82

 ç¥ŀ

-0.81

Zone

-0.80

inator

-0.79

ktop

-0.78

patch

-0.77

ĸļ

-0.77

xtap

-0.76

sembly

-0.76

Interstitial

-0.75

POSITIVE LOGITS

 wrongdoing

1.17

 collusion

1.05

 Saddam

0.99

 misconduct

0.97

 UFOs

0.96

 tampering

0.94

 corruption

0.92

 criminality

0.88

 fals

0.87

 fraud

0.87

Activations Density 0.284%

terms related to investigation and evidence, including statements regarding the absence or presence of evidence, concrete knowledge, and findings from investigations

No Comments

No Known Activations