INDEX

Explanations

phrases related to a direction or reduction, such as the word "down" at various levels of intensity

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

archives

-0.83

achu

-0.70

itia

-0.66

icles

-0.64

andan

-0.64

itive

-0.63

rament

-0.63

¶ħ

-0.62

ificent

-0.61

andise

-0.61

POSITIVE LOGITS

LOAD

1.20

graded

1.14

stairs

1.01

grading

0.95

hill

0.90

pour

0.82

loaded

0.81

 stairs

0.81

played

0.79

grades

0.78

Activations Density 2.275%

phrases related to a direction or reduction, such as the word "down" at various levels of intensity

No Comments

No Known Activations