INDEX

Explanations

mentions of security breaches and cyber attacks

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_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

HI

-0.68

HEAD

-0.65

Tue

-0.63

rison

-0.60

ackle

-0.57

GU

-0.57

odder

-0.56

HAM

-0.56

hips

-0.55

Ã©n

-0.55

POSITIVE LOGITS

the

0.92

the

0.81

 Corpus

0.77

rontal

0.74

});

0.67

 Clyde

0.67

 Royale

0.64

 Trog

0.64

Õ

0.63

Uri

0.61

Activations Density 1.280%

mentions of security breaches and cyber attacks

No Comments

No Known Activations