INDEX

Explanations

concepts related to security and safety

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

oric

-0.83

yss

-0.81

igible

-0.80

kson

-0.80

coni

-0.78

hes

-0.76

erer

-0.76

erers

-0.75

sonian

-0.75

xual

-0.73

POSITIVE LOGITS

 advis

0.93

 measures

0.92

 clearance

0.89

policy

0.89

 checkpoints

0.89

 enforcement

0.88

 precautions

0.88

 checkpoint

0.86

 breaches

0.85

 considerations

0.85

Activations Density 1.939%

concepts related to security and safety

No Comments

No Known Activations