INDEX

Explanations

phrases that indicate an increase or addition

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

æ©

-0.82

borg

-0.72

ļé

-0.69

 Classification

-0.67

è£

-0.66

\\\\\\\\

-0.66

boxing

-0.65

xtap

-0.65

Ń·

-0.65

ãĤ´ãĥ³

-0.64

POSITIVE LOGITS

 than

0.84

 mature

0.76

 realistic

0.75

 importantly

0.74

 fortunate

0.74

 prevalent

0.74

HUD

0.72

 educated

0.71

 frequent

0.71

 interesting

0.71

Activations Density 0.011%

phrases that indicate an increase or addition

No Comments

No Known Activations