INDEX

Explanations

terms and phrases that suggest things are not harmful or are considered safe

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

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

raped

-0.76

KER

-0.74

uesday

-0.67

hod

-0.67

aeda

-0.64

yi

-0.64

rain

-0.64

acked

-0.63

icken

-0.62

 Remastered

-0.62

POSITIVE LOGITS

 bystand

0.94

 harmless

0.85

 bystanders

0.83

ality

0.74

 prank

0.73

 benign

0.73

 nuisance

0.72

 sounding

0.72

 innocuous

0.72

 mischief

0.71

Activations Density 0.009%

terms and phrases that suggest things are not harmful or are considered safe

No Comments

No Known Activations