INDEX

Explanations

words related to comparative phrases, specifically indicating the order or distinction between two concepts or entities

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

bows

-0.79

Cola

-0.71

rawdownloadcloneembedreportprint

-0.68

ashes

-0.68

Maps

-0.66

 Cards

-0.65

opian

-0.65

oku

-0.65

-0.64

arij

-0.62

POSITIVE LOGITS

mentioned

0.96

 stages

0.81

liest

0.76

nder

0.75

most

0.73

 half

0.72

 part

0.72

worldly

0.71

 category

0.69

 scenario

0.67

Activations Density 0.014%

words related to comparative phrases, specifically indicating the order or distinction between two concepts or entities

No Comments

No Known Activations