INDEX

Explanations

information related to events or activities and providing instructions or details about them

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

oux

-0.83

elled

-0.82

hack

-0.82

iotic

-0.80

illet

-0.79

ulous

-0.77

inated

-0.77

aer

-0.76

ettle

-0.76

ophobia

-0.75

POSITIVE LOGITS

 maintaining

1.20

 retaining

1.16

 preserving

1.14

 awaiting

1.12

 simultaneously

1.05

 researching

1.05

 ensuring

1.00

 avoiding

1.00

 minimizing

0.98

 acknowledging

0.98

Activations Density 1.031%

information related to events or activities and providing instructions or details about them

No Comments

No Known Activations