INDEX

Explanations

information related to activities or events happening alongside a main event or function

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

oux

-0.87

iotic

-0.82

elled

-0.82

aer

-0.81

illet

-0.80

ulous

-0.78

inated

-0.77

iola

-0.77

ettle

-0.76

hack

-0.75

POSITIVE LOGITS

 maintaining

1.21

 retaining

1.15

 preserving

1.12

 awaiting

1.09

 researching

1.07

 simultaneously

1.02

 avoiding

1.01

 acknowledging

0.99

 minimizing

0.97

 keeping

0.96

Activations Density 0.463%

information related to activities or events happening alongside a main event or function

No Comments

No Known Activations