INDEX

Explanations

verbs indicating change or development

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

yet

-0.75

initely

-0.73

psey

-0.70

 respectively

-0.69

 generally

-0.68

 preserved

-0.67

matically

-0.67

 preserves

-0.66

Generally

-0.64

 inherently

-0.64

POSITIVE LOGITS

mega

0.74

rier

0.73

ounters

0.72

inders

0.69

ORED

0.69

 NETWORK

0.69

ORN

0.69

ackle

0.67

VALUE

0.65

Ô

0.65

Activations Density 0.121%

verbs indicating change or development

No Comments

No Known Activations