INDEX

Explanations

terms related to management positions or executives

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

bane

-0.79

ologies

-0.79

INGTON

-0.78

cale

-0.78

abouts

-0.76

ulia

-0.76

lust

-0.75

ERO

-0.75

atoon

-0.73

Downloadha

-0.71

POSITIVE LOGITS

 summary

0.99

 director

0.94

 branch

0.91

 producer

0.91

 committee

0.87

 officer

0.87

utive

0.85

 chef

0.84

ctive

0.83

 chairman

0.79

Activations Density 0.437%

terms related to management positions or executives

No Comments

No Known Activations