INDEX

Explanations

references to human-related topics or issues

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

RET

-0.72

 Buckingham

-0.70

href

-0.68

 Transcript

-0.67

urations

-0.65

UGE

-0.64

 Comprehensive

-0.63

eryl

-0.63

arella

-0.62

REL

-0.62

POSITIVE LOGITS

 beings

1.41

itarian

1.23

itar

1.15

oids

1.12

itary

1.00

oid

0.97

zee

0.94

 readable

0.94

 embryonic

0.94

istic

0.92

Activations Density 1.176%

references to human-related topics or issues

No Comments

No Known Activations