INDEX

Explanations

mentions of specific names or individuals, possibly related to news articles or stories

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

Els

-0.70

 Dominion

-0.68

Load

-0.67

 Agric

-0.65

ories

-0.64

 Pixie

-0.64

oard

-0.63

Native

-0.63

Georg

-0.62

 Krish

-0.60

POSITIVE LOGITS

Jr

1.19

Sr

0.95

agher

0.88

III

0.87

 himself

0.86

 famously

0.86

 confid

0.85

 joked

0.84

Jr

0.83

ovich

0.83

Activations Density 1.068%

mentions of specific names or individuals, possibly related to news articles or stories

No Comments

No Known Activations