INDEX

Explanations

adverbs indicating distance or extent

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

 Picks

-0.69

piracy

-0.68

icks

-0.65

 Peel

-0.64

 PROGRAM

-0.64

ellation

-0.63

emis

-0.63

amination

-0.58

cffff

-0.58

 Bean

-0.58

POSITIVE LOGITS

fetched

1.04

thing

1.01

aday

0.98

med

0.89

away

0.84

 flung

0.83

rior

0.79

utherford

0.78

adic

0.77

rier

0.76

Activations Density 1.074%

adverbs indicating distance or extent

No Comments

No Known Activations