INDEX

Explanations

geographical locations and political or military terms

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

owe

-0.76

urai

-0.66

 Doctrine

-0.63

talk

-0.61

orr

-0.61

ipeg

-0.56

antage

-0.56

acks

-0.56

oos

-0.55

 vantage

-0.54

POSITIVE LOGITS

by

1.23

BY

1.06

by

1.03

By

0.90

bys

0.90

ĸļ

0.89

 aback

0.84

By

0.84

 chiefly

0.81

 solely

0.79

Activations Density 1.080%

geographical locations and political or military terms

No Comments

No Known Activations