INDEX

Explanations

quantitative data related to rates

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

emis

-0.86

IRT

-0.78

worth

-0.76

paio

-0.69

iami

-0.68

ADRA

-0.68

bley

-0.67

NetMessage

-0.67

stanbul

-0.67

 Jakarta

-0.67

POSITIVE LOGITS

payers

1.48

payer

1.26

(%)

0.91

 hike

0.80

 variability

0.79

 fluct

0.77

OA

0.76

 ratios

0.76

uation

0.76

 fluctuations

0.76

Activations Density 0.588%

quantitative data related to rates

No Comments

No Known Activations